EPA-600/3-77-044
May 1977
FORMATION OF PHOTOCHEMICAL AEROSOLS
by
Edgar R. Stephens and Monty A. Price
Statewide Air Pollution Research Center
University of California
Riverside, California 92521
Grant No. 800868
Project Officer
Joseph J. Bufalini
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
LIBR
v. c. r,;vK,,.. .-;;TAL PKCTECTION AGEKg$
tC-"/.;.1. , J. G88I7
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences 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.
ii
-------
ABSTRACT
Although aerosol is the most readily apparent symptom of photochemical
smog because of its affect on visibility, its scientific study was long
neglected in favor of the more pressing questions of eye irritation, plant
damage, and health hazards. The early reports by Haagen-Smit suggested that
visible haze might be due to sulfuric acid formed by atmospheric oxidation
of sulfur dioxide but he also noted that the rapid reaction of ozone with
olefin (at high concentrations) produced a dense "fog." Since this reaction
system also caused eye irritation and plant damage it seemed fair to regard
it as a good synthetic or laboratory smog. The relative importance of the
sulfuric acid and ozone/olefin aerosols was not studied for many years nor
has there been a good quantitative, realistic simulation in the laboratory
of smog aerosol formation.
The present project had as its general objective a better understanding
of the smog aerosol with particular reference to haze as it is observed in
inland (receptor) areas of Southern California. It combined laboratory
studies with ambient air studies in the belief that full understanding
requires a reconciliation of the two. Counting of particles by light scatter-
ing was the principle physical technique while infrared analysis was the
major source of chemical information. A new technique of reflectance spec-
troscopy was developed for this purpose. Spectra were recorded by reflectance
through samples deposited by impaction on noble metal targets. This technique
requires small samples and has the unique advantage of distinguishing between
sulfuric acid and other sulfates since the sample has no opportunity of
reacting with anything but itself. This new technique is regarded as one of
the most important accomplishments of this project.
Infrared spectra of ambient aerosols have bands assigned to sulfate,
nitrate, ammonium and water which are completely removed by water washing of
iii
-------
the sample but not by benzene washing. A synthetic aerosol generated by mix-
ing ammonia with sulfuric and nitric acids produced a very similar spectrum.
One pair of ambient samples gave spectra which looked more like sulfuric acid
than sulfates. The larger particles gave a 10 micrometer band, insoluble in
water and benzene, which is attributed to soil dust. Few samples showed any
CH band of organic material and no carbonyl band was identifiable. Synthetic
organic aerosols either from ozone/olefin or olefin/nitrogen oxide reactants
do show such bands but they are not water soluble. It also appears difficult
to produce significant amounts of organic aerosol from realistic levels of
hydrocarbon/NOx mixtures. It was concluded that a much more sophisticated
apparatus would be necessary to attain a good simulation of photochemical
aerosol formation. Long reaction times at low reactant concentrations are
required and the rate of loss of aerosol to the wall must be minimized. This
calls for a small surface-to-volume ratio. While it is clear that organic
material makes a contribution to aerosol it is still not possible to uniquely
characterize it as to source and history or in its contribution to visible
haze.
Many ambient aerosol particles are hygroscopic or deliquescent so that
they swell as the relative humidity increases and shrink as it decreases.
This readily explains the heavy haze often seen when humidity is high but
oxidant is low as in the early morning hours. Frequently the visibility
improves as the temperature rises and the relative humidity falls during the
forenoon.
A variety of reaction products (SO^ , NC>3 , organics) can potentially
fulfill the requirement of water solubility and hygroscopicity (or deliques-
cence) to contribute to haze and loss of visibility. The relative importance
to haze of these species may well be different at different times and places.
Also still not settled are the exact sequence of reactions which produce these
aerosol products. These are matters of importance in control strategy and not
just of scientific curiosity. Abatement of aerosol might be achieved either
by control of precursors (for example, sulfur in fuels) or control of the oxi-
dation process which produces aerosol products. At least it has become clear
that control of direct aerosol emissions will not markedly improve visibility
in inland Southern California.
IV
-------
CONTENTS
Abstract ill
Figures vii
Tables x
Acknowledgement xi
1. Introduction 1
2. Methods 3
Particle counting 3
Infrared analysis 4
Spectraphone 11
Generator of synthetic aerosols 11
3. Results 19
Impactor sampling 29
Southern California survey 52
Extraction of organic materials 53
Photochemical formation of aerosols 58
4. Future Potential 62
References 64
Bibliography 65
v
-------
FIGURES
Number Page
1 Correction of Martens and Kellner^ ' light scattering
function for index of refraction of 1.6 for polystyrene to
sulfuric acid index of 1.4 5
2 Aerosol impactor which produces a spot on either an infrared
transparent target or a corrosion resistant metal plate 6
3 Reflectance attachment which places a 3:1 reduced image of
the source on the sample (which is vertical rather than
horizontal as shown here) 8
4 Arrangement for estimating the two-stage impactor's efficiency
in collecting ambient aerosol 9
5 Tube flow photoreactor 12
6 Dilution apparatus for sampling from the tube flow
photoreactor 15
7 Stirred flow photoreactor 16
8 Apparatus for generation of aerosols for impaction to make
reference spectra 17
9 Atomizer section of particle generator 18
10 Aerosol size spectra (cummulative) for four distinctly
different weather conditions 20
11 Concentrations of pollutants on a smog-free day in 1956 showing
accumulation of unphotolyzed pollutants under a radiation
inversion(8) 22
12 Clean air at 7:30 AM, 10 February 1971, over central Riverside.
See particle count in Figure 10 and hydrocarbon analysis in
Figure 13 23
13 Hydrocarbon analysis in "clean" air 10 February 1971 (see
Figures 10 and 12; also ref. 11) 24
VI
-------
FIGURES (continued)
Number
14 Infrared spectrum on BaF2 (large particles; diameter - ~l micro-
meter) of "clean" morning air of 10 February 1971. Note
scattering of short wavelengths and weak band at 10 microns .... 25
15 Infrared spectrum on BaF2 (small particles; diameter < ~1 micro-
meter) of "clean" morning air of 10 February 1971. The
significance of the bands at 7-7.5 micrometers is unknown 26
16 Photomicrograph of large particles whose spectrum is shown in
Figure 14. Tan color predominates; looks like sand 27
17 Photomicrograph of small particles whose spectrum is shown in
Figure 15. Black particles look like carbon 28
18 Typical smog aerosol reflectance spectrum. Band assignments in
the text. Water washing removed all bands except air bath water
and C02 and scattering at short wavelengths 31
19 Fourier transform spectrum of Riverside aerosol small particles
on KRS-5 (transmission) (run at EPA laboratories by B. W. Gay,
Jr.). Compare small particle spectrum with Figure 21 32
20 Fourier transform spectrum of Riverside aerosol large particles
on KRS-5 (transmission) (run at EPA laboratories by B. W. Gay,
Jr.). Compare large particle spectrum with Figure 22 33
21 Prism spectrum of ambient aerosol (small particles) on KRS-5
shows the same bands as the FTS spectrum of Figure 19 34
22 Prism spectrum of ambient aerosol (large particles) impacted on
KRS-5 shows the same bands as the FTS spectrum of Figure 20 .... 35
23 Ambient aerosol large particles (16 April 1973) showing enhance-
ment of some bands in response to moist air exposure. No effect
of dry air 36
24 Washing of large particle sample of Figure 23 with benzene did
not change the spectrum significantly. Washing with water
removed all but the broad 10 micrometer band attributed to
silicate 37
25 Ambient aerosol small particles (16 April 1973) showing enhance-
ment of some bands in moist air exposure. No effect of dry
air 38
vii
-------
FIGURES (continued)
Number Page
26 Washing of the small particle sample of Figure 25 with benzene
had no effect on the spectrum, but water washing removed the
sample completely 39
27 Impaction of morning air small particles. No absorption bands
are seen even though a black sample was visible 41
28 Impaction of morning air large particles. Only the broad "dust"
band at 10 micrometers is visible. Visually the sample was
light brown 42
29 Infrared spectrum of liquid water in a thin film between KRS-5
plates 43
30 Infrared spectrum of aqueous sodium sulfate between KRS-5
plates 44
31 Infrared spectrum of aqueous sodium nitrate between KRS-5
plates 45
32 Infrared spectrum of aqueous ammonium chloride between KRS-5
plates 46
33 Impacted aerosol synthesized from ammonia with sulfuric and
nitric acids. Note the close resemblance to the ambient
spectrum of Figures 25 and 26 48
34 Impacted sulfuric acid aerosol. Note three broad bands between
8-12 micrometers 49
35 Sample taken on UCLA campus. Note similarity to sulfuric acid
sample. These large particles show the dust band after water
wash 50
36 Sample taken on UCLA campus. This also shows the bands
tentatively identified as sulfuric acid 51
37 Sample of ambient aerosol mixed with high concentration ozone/
olefin reaction products. Water washing removed the natural
aerosol to leave a clear CH and carbonyl band 54
Vlll
-------
TABLES
Number Page
1 Aerosol formation from auto exhaust by photochemical formation
of sulfuric acid 2
2 Impaction efficiency of the impactor at various slit widths,
spacing and flow rates 7
3 Sulfuric acid aerosol in tube photoreactor 13
4 Infrared bands (micrometers) of ambient and exhaust aerosols ... 56
5 Carbon/hydrogen analysis of ambient aerosol extracts 57
6 Aerosol produced by the photochemical reaction of nitrogen
dioxide and ot-pinene 59
ix
-------
ACKNOWLEDGEMENTS
Several staff members of the Statewide Air Pollution Research Center
made contributions to these studies. These include Frank Burleson, George
Doyle, Glen Vogelaar, and William Snider. We thank also Dr. Philip L. Hanst
and Mr. Bruce Gay of the Environmental Protection Agency laboratories at
Research Triangle Park, North Carolina, for running Fourier transform spectra.
x
-------
SECTION 1
INTRODUCTION
Development of a complete understanding of the aerosol component of
photochemical smog has attracted increasing attention in recent years. In
addition to its health hazards, aerosol also causes a presistent loss of
visibility which is evident to every citizen. In inland Southern California
visibility is often reduced to a few miles even when the humidity is only
30-40%. Visibility impairment is quite evident even when smog is light as
judged by other symptoms. This project had the objective of clarifying the
nature of this smog manifestation with particular reference to the aerosol
which occurs in Riverside, California. The first results were described in
two papersa short one in 1970 and a longer one in 1972. It was shown
that "unreacted" air polluted with car exhaust causes very little visibility
loss even though there are large numbers of countable particles.
It was argued in ref. (2) that atmospheric oxidation of sulfur dioxide
could produce substantial visibility loss since sulfuric acid would condense
to produce particles of a size most effective for scattering light. This is
certainly not a new idea; the earliest of Haagen-Smit's papers suggest it
(3)
(without detailed proof). Even the sulfur dioxide derived from auto
exhaust would produce substantial visibility loss if a large portion of it
were photooxidized to sulfuric acid as shown in Table 1. The high oxidant
levels of inland Southern California show that the photochemical reaction is
more advanced in those areas than in most places. Enhanced photochemical
aerosol can therefore be expected. Laboratory studies have shown that sulfur
dioxide can be oxidized to sulfuric acid under simulated atmospheric
conditions (1, 2 and references therein), but some hydrocarbon/nitrogen
oxide mixtures will produce an organic aerosol in the absence of sulfur
dioxide. It has proven difficult to determine the relative contributions
of these two distinct types of aerosols in real polluted air. The problem is
-------
TABLE 1. AEROSOL FORMATION FROM AUTO EXHAUST BY PHOTOCHEMICAL FORMATION OF
SULFURIC ACID
1.
2.
3.
Gasoline
combustion
Exhaust gas
dilution
Polluted air
photooxidation
v 15
v 10,000
X 1
0.
0.
0.
06%
004% -
004 ppm
0.06% sulfur by weight
15 grams exhaust/gram fuel
40 ppm S by weight =
40 ppm SO 2 by volume
S02
Assume 100% conversion
with HC/NOX
4. Sulfuric acid
absorption of
moisture at
40% RH
5. Smog aerosol
X 2
0.004 ppm H2SOtt = 16 micrograms/meter3
32 micrograms of 50% H2SOi+/meter3
360 million droplets
(D 0.3 - 1.0 micron)/meter3
complicated by the fact that the respective weights of various aerosol compo-
nents are not necessarily a reliable guide to the contributions to visibility
loss. Many different kinds of carbon, both inorganic and organic (even bio-
logical) can contribute to aerosol weight so that measurement of "benzene
soluble" or total organic carbon still does not quantify the contribution to
light scattering nor can it distinguish between primary organic aerosol (for
which there are many sources) and photochemically formed organic aerosol.
This field of study has been very active as is shown by the list of recent
references dealing with some aspects of aerosol as related to visibility (see
bibliography).
Since Riverside has frequent episodes of medium to heavy smog including
heavy haze it is an ideal place for this study. A primary aim was to use the
same techniques on both natural and synthetic aerosols. One of the difficult
problems, not only in aerosol studies but in any laboratory study of smog, is
in establishing the importance of any laboratory finding in the real atmo-
sphere. This gap can only be bridged by achieving quantitative and consistent
results in both settings.
-------
SECTION 2
METHODS
The basis of this aerosol study was to apply the same aerosol measure-
ment techniques to ambient air in Riverside and to aerosols generated in the
laboratory. This approach is potentially much more informative than either
kind of study alone. The aim is to generate in the laboratory an aerosol
which matches as closely as possible the properties of the ambient aerosol.
Two principal methods of aerosol analysis were used in this study, one to
provide physical information and one to provide chemical information.
PARTICLE COUNTING
Particle size distributions were measured with a Bausch and Lomb 40-1A
forward-scattering particle counter. This can be used to count particles
larger than 0.3 p (polystyrene equivalent) diameter. As supplied by the
manufacturer, this counter has a lower-limit discriminator which can be set
to any of eight levels. It then provides a count of total particles larger
than the given diameter. Thus, a size spectrum can be obtained by switching
the counter from one minimum diameter setting to another. This has a disad-
vantage, especially when working with ambient samples, that the count for
each size is made on a different sample of air, and there is no way to be
certain that the particle size distribution remains constant for the approxi-
mately 16 min required to measure one spectrum. Another disadvantage of this
counting system is that when measuring particles of large size, which are
very few in number, all the information about the smaller diameter particles
is being wasted. Within these limitations, this counter appears to be an
accurate and reliable means of estimating the number of particles in various
size ranges in laboratory aerosols and in ambient aerosols. It requires a
rather small flow rate which is one of its major advantages for use in labora-
tory studies, since laboratory reactors provide limited volumetric flow. The
-------
counter is calibrated in terms of polystyrene equivalent diameters. Whatever
the ambient aerosol is, it is certainly not polystyrene. It may much more
likely be sulfuric acid. To make a more realistic size discrimination, the
polystyrene equivalent diameters were corrected to sulfuric acid equivalent
diameters for some studies. Plots of scattered light derived from Martens
(4)
and Keller were used for this purpose as given in Figure 1. The counter
was used to count aerosols formed in the laboratory reactor, and to count
ambient aerosols under various weather and pollution conditions.
To make full use of the information provided by this counter its elec-
tronics were completely rebuilt to provide an input for a 200-channel pulse-
height analyzer. This turned out to be a rather major operation but it was
successfully completed. This permitted simultaneous measurement of particles
of all sizes which assures spectrum consistency and also permits much finer
resolution. All the data in this report were completed before this modifi-
cation was complete.
INFRARED ANALYSIS
To provide chemical information on these synthetic and natural aerosols
the special impactor shown in Figure 2 was developed. It provides a spot
sample of aerosol which matches the geometry of the entrance beam of a small
infrared spectrophotometer (Perkin-Elmer 137B) (15 x 1.5 mm). This permits
the recording of transmission spectra. Three versions of this scheme have
been explored. In the first version infrared transmitting targets such as
BaF2, BaF2/CaF2 (T12) and TIBr/TlI (KRS-5) were used. These crystals not
only transmit infrared but they are also insoluble in water. The KRS-5
crystal covers the 2-15 micrometer wavelength range whereas the other two
are limited at the long wavelength end. Good spectra can be obtained within
these limits although ambient air samples etch the targets to form a scar
which must be polished out after each use. This etching implies that the
samples react with the target and must themselves therefore be changed. This
is consistent with the identification of strong acids as components of the
aerosol. This result prompted the investigation of the use of noble metals
as targets combined with reflectance spectroscopy for analysis.
To obtain a good reflection spectrum it is necessary to ensure that most
of the infrared beam pass through the sample. This requires that the source
-------
q
d
q
in
o
cc
q o
CM ,-
tr
LJ
h-
UJ
Q
LJ
-I
O
ir
If) Q.
6
ro
d
00
o
o
o
CVJ
b
ro
b
X
QJ
"B
H
S-l
O
c 4-4
o o
O
C
3
6013
a -H
i-t O
5-i ctf
rt
O
03
,c en
oo
H O
x Q)
- C
0)
!-( H
QJ >.
C -u
^ o
a.
T3
C l-i
nj O
4-<
03
C o
OJ
4-1 iH
5-1
n) u-i
O O
H
a *->
o o
H nj
4J M
O 4-4
0) OJ
O 4-4
0 O
1H9I1
Q)
)-l
3
1-1
fn
-------
£
WE
0 E
£*
< x
0. 1C
cr
o
i-
o
<
Q.
ll
/
/
^ifl+tlfltif+WKt
\
<\ l
Ay
1 V
f r
/ /
\
V,
V
/,
/v
c
d)
M
n)
a,
CO
C
n)
01
S-i
tfl
H
H
(U OJ
4-1
a nj
O i-t
ex
4-1
O i-i
& n)
03 4-1
Ol
ni 6
en 4-1
01 c
o to
3 4-1
T3 CO
O -H
M M
CX 0)
M
O C
O
-H
s
H
O H
4-1 O
o o
CO
{X cd
O 4J
03 0)
O 60
M M
O) tO
< W
(1)
J-l
60
H
-------
of radiation be imaged sharply on the sample and that this image in turn be
sharply imaged in the entrance slit of the spectrometer. Results were not
satisfactory with an ordinary reflectance attachment but a microspecular
reflectance attachment (Barnes Engineering Co., Model No. 126) yielded strong
reflectance spectra (Figure 3). This accessory places a 3:1 reduced image
of the source on the reflective sample surface. This would make it possible
to use an impactor of shorter slit length which would in turn permit reduction
in the sample flow rate by a factor of three. This would be particularly
advantageous in sampling from laboratory reactors. It might permit reduction
of the sample size to 300 liters for heavy aerosols. Such a small impactor
has not yet been constructed.
Impactor efficiency was estimated by counting the particles in a sample
of air which had been passed through the impactor using the system shown in
Figure 4. A 35-liter steel tank was inserted in the line between the impactor
and the vacuum pump. In this way the tank could be flushed with impactor
effluent. The slit widths, spacings and flow rates tested are listed in
Table 2. While the tank was purging, the size distribution of the aerosol
TABLE 2. IMPACTION EFFICIENCY OF THE IMPACTOR AT VARIOUS SLIT WIDTHS,
SPACING AND FLOW RATES
% Impacted
Slit width (mm)
Spacing
(mm)
Flow rate (1/min)
Aerosol
size
y
0.3-0.5
0.5-
1.0-
1.0
1.8
1.8-2.0
2.0
+
0.76
0.25
39
27
50
64
86
100
0.76
0.25
28
35
59
69
83
100
0.76
0.51
28
45
73
84
100
100
1.3
0.25
28
39
67
71
86
100
1.3
0.51
28
-15
51
85
83
100
1.3
0.65
28
8
44
68
71
100
0.38
0.38
28
87
95
94
100
100
in the ambient air was measured with the Bausch and Lomb 40-a particle counter.
Then the pump was stopped and the aerosol in the tank counted with the results.
For most sampling the first stage had a slit 1.3 mm wide at a target distance
-------
cc
O (£>
< OJ
< Q
O O
I- ^
Q_
O
UJ
O
O
UJ
cc
a:
o
UJ
O
cr
o
s
o
UJ
a.
co
a) a)
o M
u a)
3 J=
O
cn c
^
QJ O
IM cn
O cfl
a) rH
bO rt
n) 4J
H O
N
T3 -H
a) v-i
o o
3 ^
T3
QJ C
CO !-i
a)
n) -C
4J
w n)
a) M
o
CD rH
cu
>
cn
-rl
c
OJ
e
f
U
03
nj 01
^H
cu p*
o B
C a)
n) cn
4J
O cu
cu ,c
cl
o
m
cu
60
H
-------
u
c
0)
cr
UJ
10
ro
LdCC
-JUJ
28
cc
o
o
U-l
14-4
0)
V-l
o
4-1
O
(fl
01
60
n)
4-1
en
§
0)
4J
60 iH
C O
H
4-1
CD
O
03 M
e CD
H cfl
4J
W 4J
Q) C
QJ
VJ -H
O JO
M-I e
ni
4J
a M
QJ (3
0 -H
Ol 4-1
60 O
C 0)
CO rH
I-J rH
n o
-------
of 0.65 mm while the second stage had a width of 0.4 mm and a spacing of
0.4 mm. Balanced against its obvious shortcomings this procedure has the
advantage of referring the test to the sample, ambient air, of immediate
interest.
Most of the best data were obtained using a two-stage impactor designed
to cut the sample at about 1 or 2 microns diameter. This should put the
windblown dust on the first stage and most of the photochemically generated
particulate on the second stage. A flow rate of 28 liters/min was selected.
This gives a strong spectra on both stages of the impactor in two hours
sampling for moderate smog conditions. Shorter sampling times can be used
for heavier ambient aerosols. Hundreds of spectra were obtained with this
system.
To provide additional information, procedures were devised for treating
the samples with moist air or dry air while recording the spectrum. Samples
were also washed with various solvents taking care to remove the sample only
by solution rather than by mechanical flushing. The spectrum could then be
recorded again.
Two additional techniques of aerosol sampling for infrared examination
were tested during the course of this project. These were thermal precipi-
tation and the LEAP spinning disk sampler. The thermal precipitator oper-
ates at a very low flow rate and therefore requires a long sampling time.
It has the advantage of collecting very small particles. Even twenty-four-
hour samples did not suffice to collect a sample adequate for infrared
analysis, however. The LEAP sampler handles a very large flow of air but
produces an aqueous solution or suspension of the aerosol. Trial experiments
with this sampler were not very informative.
One other program of infrared study of aerosol samples has been carried
out at Argonne National Laboratories by Cunningham and associates. ' They
collected Lundgren Impactor samples which were transferred into KBr powder
and pressed into pellets for infrared examination. Although Fourier transform
spectroscopy was used, their published spectra did not use high resolution,
probably because no fine structure was present in the spectrum. Generally
speaking, these spectra are similar to the ones generated in the present
project and were interpretated in terms of sulfate/nitrate with little
10
-------
evidence of organic CH bands. Aside from the extra procedure of preparation
of the pellet this method has the disadvantage of allowing interaction of
sample with the salt (as compared to the reflection method). The advantage
is in the potential for a quantitative analysis.
SPECTRAPHONE
Although several different methods of aerosol sampling have been used
in this and other projects no one of them is ideal. A system which would
permit analysis of the aerosol phase without separation from the gas would
eliminate many uncertainties in the sampling process. Accordingly, schemes
for recording the spectrum of the aerosol phase directly have been explored
in considerable detail. The spectraphone concept, based on heating of a gas
by absorption of radiation in the characteristic wavelength bands, was
investigated in a detailed paper study. Since gases as well as particulates
absorb radiation, a substraction technique must be used to extract the desired
spectral information. The details of the design analysis are being given in
another report (in progress).
GENERATION OF SYNTHETIC AEROSOLS
Several different methods for generating synthetic aerosols and obtaining
their spectra were used during the course of this project. These ranged from
a laboratory flow tube reaction to vehicle engines of various types. Two
types of "atomizers" were also used to make aerosols for spectra and some
spectra of inorganic aqueous solutions were recorded by conventional tech-
niques .
Photoreactors
A tubular flow reactor, which has been used for a number of years to
study simulated photochemical reactions, was used for this study. It is
made from sections of borosilicate glass drain line 15.24 cm (6 in) in diame-
ter in the form of a U-tube, each arm of which is about 3.05 meters (10 ft)
long (see Figure 5). The total volume is about 140 liters. With a flow rate
of 2.35 liters/min the residence time in the flow reactor is 1 hr. The humid-
ity in the incoming air was maintained at a standard level with an ice bath
humidifier. Pollutants are then added at a low concentration and the mixture
11
-------
1
i
o
-------
passed into the reactor. Each 5-ft section of the reactor is surmounted by
six blacklight fluorescent lamps to provide simulated sunlight. The effluent
from the reactor can be analyzed for chemical products or for aerosol.
In addition to aerosols generated chemically either by photolysis of
hydrocarbon/nitrogen oxides/sulfur oxides mixtures or hydrocarbon/nitrogen
oxides mixtures, synthetic inorganic aerosols were also made in the tube
photoreactor. Sulfuric acid aerosol was prepared by vaporizing fuming
sulfuric acid into a stream of air. A particle size distribution obtained
for a typical synthetic sulfuric acid aerosol as it entered the tube reactor
and as it left the tube reactor are shown in Table 3. From these data it
TABLE 3. SULFURIC ACID AEROSOL IN TUBE PHOTOREACTOR
Particles/ft3
Diameter Input Output
Microns
D I
0.3 6.397 x 106 6.004 x 106
0.5 1.430 x 106 1.414 x io6
1.0 0.254 x io6 0.236 x io6
1.8 0 0
Calculated yg/m3
1.0
1.8
> D
> D
> 0.3 11.7
i 1.0 5.3
Equivalent ppb SOX
1.0
1.8
> D
> D
Measured
> 0.3 2.9
> 1.0 1.4
SOX ppb 620 640
appears that there was very little loss of aerosol in the one hour time of
residence in the flow tube. For some unknown reason the conductivity analyzer
indicated a far higher t^SOi^ concentration than the particle count. We can-
not, of course, be sure that the particles leaving the reactor were the same
as those entering the reactor since only a portion of the total size spectrum
is counted and particles too small to be measured in the reactor input may
well have coagulated in passage through the tube, and larger particles may
13
-------
have fallen out during passage through the tube. Sampling from this reactor
with the impactors presented some difficulty since the flow from the reactor
was less than one tenth of that needed for impaction. The scheme shown in
Figure 6 was developed to solve this problem. The reactor effluent was
diluted with filtered compressed air and the diluted aerosol passed through
the impactor. A stirred flow bottle reactor shown in Figure 7 was also used
for some experiments.
Aerosol Generators
To generate test aerosols for impaction spectra the nebulizer shown in
Figures 8 and 9 was constructed. It produces a stream of airborne particles
from either a polystyrene latex emulsion or from a water solution of soluble
salts. Since the ambient aerosol has a large component of aqueous droplets,
some reference spectra were also obtained by pressing water solutions between
KRS-5 plates. The water absorptions are so strong that no spacer could be
used, only the residual film between the two infrared windows.
Auto exhaust particulate was sampled using a dilution system similar to
that shown in Figure 6. These experiments were qualitative and only intended
to provide reference spectra for the ambient air samples.
14
-------
M
O
4-J
O
cfl
0)
O
ex
3
4J
0)
§
t-l
00
H
r-H
&
4-1
cfl
a.
cfl
a
o
rl
Q
60
H
15
-------
M
O
4-J
U
n)
0)
M
O
4-1
O
ex
OJ
S-i
H
4-1
C/3
ai
M
60
H
-> A CC
t/> f _i&
£ °=£
-- '
-------
0)
a
c
0)
a:
I
£T
UJ
UJ
UJ
o
cr
2
o
4-J
c
o
H
4-1
O
cfl
a.
S
M
O
4-1
O
(0
o
t-l
0)
CT3
C
o
H
4-1
cfl
M
(U
C
0)
O
M-i
to
3
4-1 CO
Cfl J-l
H 4J
Cfl O
P. 0)
a, a.
-------
'AIR INLET TUBE
^SPLASH GUARD
VENTURI ATOMIZER
Figure 9. Atomizer section of particle generator.
18
-------
SECTION 3
RESULTS
Heavy photochemical smog gives about 107 particles/ft3 (350 particles/cm3)
countable with the Bausch and Lomb instrument (larger than 0.3 micrometer
diameter). The number of course, falls off steeply with increasing diameter
as shown in Figure 10 which compares a day with a very high oxidant (0.65 ppm)
to three days of distinctly different weather and with much better visibility.
In each of these cases there were sufficient particles to provide condensation
nuclei. One of the commonest questions posed by laymen about smog aerosol
is "How much of this is smog and how much is fog?" or, when the oxidant and
sunshine are low but the visibility is impaired, "But this isn't smog is it,
isn't it just haze?" The answer seems to be that the aerosol from smog
persists even when the oxidant is low or after the oxidant has been destroyed
by NO emissions. Cases can be cited of equivalent or higher humidity but
much better visibility in Riverside and elsewhere. The "haze" which persists
when oxidant and photochemistry are minimal must be due to residual pollution
which can impair visibility noticeably, especially if the pollutants are
hygroscopic and if the relative humidity is high. They cannot be explained
by lack of condensation nuclei; these are never in short supply in the real
atmosphere. But high humidity alone will not produce a visible haze. The
lowest curve in Figure 10 was taken on an overcast, rainy day when the
humidity was 78% (much higher than typical smog) (Temp. 61°F, wet bulb 57°F)
but the visibility was unimpaired. Any haze due to natural causes should
have been visibile on such a day.
Santa Ana wind episodes produce some of the best visibility this area
ever enjoys. Although under extreme conditions these winds from the north
and northeast can create a very heavy loading of sand and dust, milder winds
produce lower counts as shown in Figure 10. A few times each year, especially
in winter, a mild Santa Ana condition is followed by a calm, cloudless night
19
-------
c
0)
H
CU
O
C
CO
H
-d
M
3
O
14-1
0)
H
4J
cfl
d
a
rt
!-J
4-J 03
o C
(U O
ft -H
en -u
H
0) T)
N a
H O
03
CO XI
O 4J
Vj nJ
-------
which produces a strong but shallow radiation inversion at dawn. If radiation
cooling of the surface and its vegetation are strong enough, orchard heaters
will be ignited to protect the citrus crop. At dawn and through the morning
hours the smoke from these heaters produces a black pall which hangs over
the valley. This aerosol not only appears black in the sky (in contrast to
photochemical smog which is usually white) but it appears black when impacted
on a target. Such a sample, since it is principally elemental carbon parti-
cles, shows no infrared bands. If the minimum temperature is above freezing
the heaters will not be used but the radiation inversion is still very strong
and will trap vehicle exhaust gases from the morning traffic as well as
industrial emissions. High levels of primary pollutants (CO, NOX, EC etc.)
can be measured even though the visibility is unimpaired. This observation
was first made by P. L. Hanst in December 1956 in South Pasadena (Figure
/ Q \
11). It clearly demonstrates that unreacted auto exhaust (and with lesser
certainty other sources) makes little or no contribution to visibility loss.
Additional measurements made on similar mornings in Riverside have amply
confirmed this conclusion and provided some additional data. One such morning
was 10 Feb 1971 when early morning samples were taken on the roof of a five-
story office building in central Riverside. The photograph Figure 12 shows
unimpaired visibility of the San Gabriel mountains about 21 miles distant.
Analyses by gas liquid chromatography, in contrast, show high levels of hydro-
carbons in the pattern of auto exhaust emissions (Figure 13). The particle
count, included in Figure 10, was about ten times lower than for heavy smog
although the hydrocarbon levels were several times larger in this early
morning sample. Even this particle count is larger than would have been
anticipated from the apparent visibility. Perhaps these observations could
be reconciled by suggesting that those particles present in the early morning
are black rather than white. Impaction samples taken at this time showed
only tiny infrared bands (Figures 14 and 15), but the sample was visible on
the BaF2 target. In fact, photomicrographs showed these particles (Figures
16 and 17).
This observation is dealt with at some length because it has a direct
practical application to control strategy: control of primary auto exhaust
particulates will not result in any visibility improvement. On the other
21
-------
1
1
3
o
o
00
c
H G
O -H
,£ CO
cn M
CD
vO >
u~i G
>> rt
nj -H
I M
60 a)
O T3
e G
cfl W
to 3
CO O
14-1 O
O 4-*
o
cn JG
G cx
O G
H 3
4->
Cti >*-(
!-l O
C 4J
0 a)
U iH
3
60
H
22
-------
0)
CU
C/l
Q) 0)
"° bl
H d
en M
H -H
0) fn
03
-r-t
M
C -H
a) cd
o C
-
a) c
> o
O £>
O
O
03 T3
3 C
s-i cd
^D
-------
0)
-l
o
d .H
rt
0)
iH 14-1
a % -H
tfl fe
CD
V-i
M
H
24
-------
ro
Q
O
OJ
O
ro
^ ; if)
C5 C>
P 8
s-^ U
M V-l
01 O
4-1 &
QJ CO
6
O 4-1
SJ O
O
H 60
6 C
H
Al -U
03
!-i O
o) tn
0) 0)
f3 A-J
cfl O
13
U rH CO
H a
J-i >, O
^ t-i j-i
tfl tfl O
ft 3 -rl
M 6
0) ,£2
00 0) O
M PH rH
«
rH O -U
^ rH tfl
T3
O C
rt m
m M .Q
H
d n) ^
O cfl
i-i C T3
u VJ C
O O n)
0) B
a, 03
w = ,c
*"O f^ ^0
CU 0) C
J-l rH Q)
tfl O rH
VJ = 0)
MH >
C M-I cd
M O S
CD
»-i
M
H
25
-------
I I
I I I I I
m
cvi
I
o
'e
01
oo
UJ
_1
LU
in
ro
O : c\J ro
-------
3
60
H
fi
H
o
en
en
6
S-l
4->
o
0) t)
d, C
en nj
U)
cu
05 01
O ^
fi -H
& r-l
co en
iH O
O O
H r-l
4J
n) co
4-1
cu td
M -rl
iH O
13
m a)
o M
X
GO o
o o
}-<
o ti
H a)
8 H
O
4-1
O
(U
M
3
bO
rl
27
-------
3
60
H
STr&wVV' .r
J«%ir*-r |' *
^fe*^:
c
s
o
03
H
6
M
4J
O
a)
Pu
03
a)
U3
O
03 t-i
(U 03
i-l O
o
H
-------
hand, secondary pollutants formed by the photolysis of auto exhaust play a
vital role in generation of light scattering particulate.
IMPACTOR SAMPLING
The appearance of the photochemical smog samples from the two impaction
stages are strikingly different. The first stage (larger particles) supports
a golden brown deposit about one mm wide by 15 mm long while the smaller
particles on the second stage give jet black lines narrower than the first
stage. With moderate magnification the sample of larger particles shows
many tan, white and brown angular particles which are obviously solids with
irregular shapes. These are considered to be soil dust.
It is evident that windblown soil dust makes an important contribution
to total aerosol weight by contributing large particles (diameter greater
than 1 or 2 micrometers). The sample of smaller particles shows many black
soot-like particles. It is probable that these are in fact carbonaceous
particles from combustion sources (autos, trucks, etc.). Both samples when
viewed with magnification showed many liquid drops. When exposed to humid
air these drops grow and coalesce into a dirty liquid mass.
The model 137 spectrometer is a simple student-type prism instrument
with low resolving power. This is no handicap because the aerosol spectra
have no fine structure to be resolved. A typical spectrum of smog aerosol
obtained with the microspecular reflector is shown in Figure 18 (second
stage = small particles). The particle count at sampling time (1015 to 1050)
was 12 x 106 ft" . Heavy haze was present and a strong spectrum (too strong
actually) was obtained. The bands are assigned as follows: 3.0-3.5 microns =
primarily water (OH stretch) with ammonium ion and some CH possible (3.4
microns); 7-8 microns = ammonium and nitrate ions; 9 microns = sulfate ion;
12 and 14 microns = nitrate ion. The bands at 5 and 6 are ascribed at least
in part to liquid water. The carbonyl stretching vibration at 5.75 micro-
meters would be a useful diagnostic test for organic molecules if it could
be clearly identified. It may be noted that the band at that wavelength
(5.75) in PAN is due not to a carbonyl but to a peroxy nitrate group.
More detailed interpretation of these spectra are not easy especially since
significant shifts of the band maxima occur depending on sample conditions.
If these shifts could be interpreted unambiguously such spectra would be a
29
-------
much richer source of information. Careful washing of this sample removed
most of the spectrum. The resulting spectrum, also shown in Figure 18, shows
bands at 4.3 and 5.5 to 7.5 microns which arise from gaseous C02 and H20 in
the extra air path of the reflectance attachment. Aside from general attenu-
ation at the shortest wavelength ascribed to scattering no absorption bands
survived the water treatment. This experiment, repeated many times, is
taken to mean that the bulk of the impacted aerosol is water soluble.
Spectra were run at low resolution on a Perkin-Elmer Model 137 (NaCl prism).
Through the courtesy of Dr. P. L. Hanst and Mr. Bruce Gay one pair of samples
were run on a Fourier Multiplex Spectrometer at EPA laboratories in North
Carolina. These spectra (Figures 19 and 20) (taken by transmission through
KRS-5) show no fine structure not present in spectra of similar samples
taken with the Model 137 (Figures 21 and 22). In fact, the spectra were
nearly identical. Figures 23 through 26 show spectra obtained on a day (16
April 1973) of lower aerosol count (6.5 x 106/ft3) with two-hour sample
(1445-1645 PST). After running the spectrum with the sample exposed to
(ambient) room air it was rerun while exposed to dry air and then a third
time while exposed to wet air (Figures 23 and 25). Then the spectra were
run once more in ambient air before they were washed with benzene and then
with water (Figures 24 and 26). Neither exposure to dry air nor benzene
washing caused significant change in the spectra. Exposure to wet air
changed the spectrum substantially, enhancing (usually) the bands ascribed
to water and changing the shape and sometimes the wavelength of the maxima,
A broad band centered near 10 microns in the large particle spectrum survived
the wash by both solvents (Figure 24). This substantiates the assignment of
this band to silicate mineral dust. None of these spectra showed any sign
of a CH band after water washing. It is possible to judge the importance
of windblown dust and carbonaceous particles by sampling when these conditions
predominate. Santa Ana wind conditions provide atmospheric dust although
the loss of visibility and the aerosol count are not as large as one would
imagine (see Figure 10). When orchard heaters are used on cold nights
carbonaceous aerosols accumulate to form a visible but black aerosol. Neither
of these aerosols show any infrared bands when impacted.
It was previously reported that the aerosol present in unreacted but
contaminated air had about one tenth the particle count of photochemical
30
-------
ro
r--
cc.
a.
C\J
o
o
5
CO
LlJ
Q
CO
cc
LJ
S-i
(3 01
H 4J
CO
en 3
4-1
C -C
QJ 4-1
6 co
C &
00
H VJ
Cfl -H
en cfl
cO
4J
a ex
C 0)
CO O
« X
QJ
en ,c
6 T) 4J
3 C 00
M CO fi
4J X3 <1)
a >n
QJ iI QJ
CX rH >
cn co cfl
£
H
CO O O
4J 0 J3
a aj en
0) M
iH 4-J
4-1 00 Cfl
0) C
>-i -H 00
,C fi
iH cn -H
O cfl H
cn ;s QJ
O 4-1
M J-l 4-1
QJ QJ cfl
Cfl 4-1 O
co cn
00 13
o
CO
cfl
rH X CM
cfl OJ O
CJ 4-1 C_)
H
a Q) T)
>> ,e c
H 4-1 CO
QJ
S-l
oo
H
31
-------
-L
en
a)
o
H
M
rt
rH
o en rH
en 01 cs
O -H
l-i M CD
01 O 1-4
CO 4J 3
cO 60
CO cB &
CU -H
"H pH
Pi [ij g
m u1 j-i
O CS 4J
O
B C cu
3 3 CX
1-1 i-i en
11 ^^/
o ^ a)
0) C rH
Cu O O
en -r-t -H
en -u
B CD M
i-i -H n)
en ni rH
C 1-1 rH
cfl JJ cfl
in
!-j i a>
cu w M
H ed co
rl 2 P.
3 e
o a o
Pu O CJ
3
M
H
32
-------
33
-------
_
CD
CO
"§
oc
o
CO 5
rO
C\J
CL
LU
CO
CO
o
o
2.
CO
UJ
o
CO
a:
UJ
>
a:
8
o
fO
o o o
m N- '
03
s
o
^c
en
c
o
01
a)
iH
O CT*
H rH
J-)
S-i
-------
-Hro
C\J
N
Q_
bJ
cn
CVJ
o
O
(S)
LJ
O
(/)
CC
LJ
>
cc
c
o
0) O
4-1 CN
O
0) CD
ex s-i
B 3
H 60
H
y-v fn
CO
-l
4-1 CO
O ^
(!) O
a, f,
cn cn
6 ^i
CO I
H CO
CN
CN
0)
M
35
-------
e ^
0) (-1
e -o
0)
O 4-1
c o
C O
(U 0)
m
MU-i
C OJ
H
5 o
O 55
,C
eo
tn
O
X
JJ
S-i 0)
n) cn
fx C
o
C B
0) o
H cn
C-J
cu
3
60
36
-------
O iI
C co
T3 T)
H 0)
T) >
O
%
0) rH
H 4-1
a, c
cfl u
en -H ^
IH OJ
-H
C
O M S
H -H O
0)
Q)
M a
nj £ *H
ex 3 B
0) 4J O
60 O iH
s-i cu
CO CL, T3
i-H en cfl
O
4-(
-------
c
0)
B
cu
a
C
03
^3
C l-i
(1) -H
n)
00
C >-,
H M
5 -a
o
^a M-I
en O
^-v 4-J
m o
r-~ a)
c> 4-1
r-H <4-l
Q)
I-)
H O
!-i Z
a,
a)
1-1
en
cr
o
en
o
cu O.
-H X
O 0)
T-t
[ 1 Ij
S-J -H
03 03
4-1
i-H en
TI -H
o3 O
e e
U5
c
rH -H
O
en en
O T3
0) 03
4-1 CD
c e
0) O
H en
m
CM
50
H
38
-------
0)
c
01 01
N X!
f. >
LTl
OJ 00
C
0) -H
oo
rl
0) |5
ex 4J
6 3
cfl 43
w
a) S
-H 3
O )-l
H 4-1
4J O
S-l QJ
cti ex
ex en
T-4. 0)
Cfl 4-1
CO C r-l
O 0)
0) 4J
XI 4-1 0)
4-1 O iH
Q) ex
4-14-10
O U-| O
a) o
60
C o 01
H a IH
xi ex
w is e
cfl cd n3
12 XI 0)
vO
CN
3,
39
-------
smog even though there is no apparent degradation of visibility. The early
morning hours of the 25th of January 1974 presented an opportunity to obtain
spectra of such aerosols. A pair of samples were taken from 7:30 to 9:30 PDT
in Central Riverside after a good atmospheric purge. Chromatographic hydro-
carbon analyses indicated high concentrations of unreacted auto exhaust; the
acetylene increased from 35.6 to 71.8 ppb during the sampling interval.
-4 -1
Visibility remained excellent and bsca(- values of 0.4 x 10 m were observed
in the area. Surprisingly the visual appearance of the samples was similar
to that of photochemical smog samples; the larger particles had a tan color
while the small particles appeared black. The black sample (small particles)
showed no infrared bands (Figure 27) while the large particles showed a broad
band at ten microns which is ascribed to silicate minerals (Figure 28). These
observations indicate that these primary aerosols do not contribute substan-
tially to visibility loss.
The assignments of these bands are based on reference spectra, one set
of which was obtained by pressing water solutions of inorganic salts between
KRS-5 windows. The bands of liquid water are so strong that no spacer could
be used. Therefore the spectra cannot be used quantitatively. Water (Figure
29) shows strong bands at 3.0 and 6.1 micrometers, analogous to the bands of
water vapor. There is a weaker band at 4.8 micrometers. Sodium sulfate
solution (Figure 30) showed, in addition to these water bands, a strong band
at 9.1 micrometers. Many ambient air spectra show a strong band near this
wavelength. Sodium nitrate solution (Figure 31) shows a strong band at 7.2-
7.3 microns assigned to the nitrate ion. The weak band at 12 micrometers
apparently coincides with the distinctive sharp band at this wavelength in
ambient aerosol spectra. The difference in sharpness is attributed to the
liquid water environment. The absence from Figure 31 of the weak band at
14 micrometers which is ascribed to nitrate in the ambient aerosol spectra
is also attributed to this environment. Ammonium chloride solution (Figure
32) shows a strong band at 7 micrometers and the water band at 3 microns is
broadened on the long wavelength side. This is, of course, the NH stretching
40
-------
CO
O
C
rt
o
H
00
S-i
o
w
,0
a)
en -H
0) en
rH -H
o >
H
4-1 M
CO
z
o
oc
o
00
in
CO
o
o
CVJ
O
to
o
m
LU
Q
CO
o:
UJ
>
QC.
rH i-H
rH 0.
ca 6
e c«
CD 05
H O
0) 0)
rH
6042
H cfl
(3
M f,
O M
6 3
O
«4H ^
O W
c c
O 0)
H >
4J 0)
c« C
a
-------
to
z
o
o:
o
oo
IT)
ro
o
CJ
O
to
o
m
o o
m
C\J
UJ
Q
CO
tr
UJ
>
cr
4-1 50
cn -H
3 -H
T3
r en
a
T3 S
CT)
O 0)
!-i -H
,£> ft
B
a) rt
,c; cn
4J
(1)
ctf
en d
a) en
rH -H
o >
cfl
rH
en
n -H
H
ni en
S-i
bo 0)
C -u
H
H
42
-------
O.
I
en
c
0)
a)
CM
O
IT
O
H
U-l
C
H
,C
4-1
tfl
C
H
J-l
(1)
4-1
Cfl
O O O
~ CO rO
O O
m N:
in
lO
en
cc
^
i
tr
LJ
ro
n
H
3
a*
H
e
O
0)
a.
CD
a)
(-1
oo
H
43
-------
o o o
m is '
30NV8dOS8V
05
0)
ca
I-H
P.
C
QJ
0)
0)
4-1
CO
3
cn
H
T3
O
CO
CO
3
O
0)
cr
cd
B
O
CU
p.
CD
M-l
c
o
01
bO
H
-------
CO
0)
4J
tfl
iH
a.
i
CO
13
0)
4-1
Q)
V
4-1
fl)
B
H
T3
O
CO
CO
O
0)
3
cr
14-1
o
S
O
Q)
CX
CO
cu
M
cfl
CD
^
3
45
-------
If)
rO
QL
CJ
UJ
CO S
DC.
Q
o
o
LO
I
en
c
Q)
(1)
0)
13
H
O
B
H
C
O
cn
p
o
a)
cr
cfl
4-1
o
B
O
(U
Cu
to
0)
S-i
03
CM
CO
0)
^
M
-H
46
-------
vibration.
Spectra of the combination salts, ammonium sulfate and ammonium nitrate,
and various salt mixtures were also run and found to be consistent with super
position of the simpler spectra. There is, however, evidence that spectra
are dependent on water content. Most striking is the behavior of the nitrate
bands; when a nitrate salt is dried, the bands at 9.7, 12.15, and 14.1 micro-
meters become much stronger and sharper. Synthetic aerosols were generated
from the sulfuric and nitric acids mixed with ammonia, and impacted onto a
reflectance plate (Figure 33). The three nitrate bands are sharp in this
spectrum. The sulfate band is unchanged in wavelength while the ammonium
band at 7.0 merges with the nitrate band at 7.35 micrometers. This spectrum
shows a quite close match with the typical ambient air sample of the smaller
diameter particles (see Figure 26 for example).
It was also possible to record spectra of sulfuric acid alone by impact-
ing a synthetic acid aerosol onto a metal target. For this spectrum (Figure
34) the strong sulfate band at 9.15 is replaced by three broad humps at 8.9,
10.0, and 11.7 micrometers. A few spectra, taken at UCLA, showed bands very
similar (but not at exactly the same wavelengths) to these (Figures 35 and
36). The nitrate bands are absent from these two UCLA spectra. The reason
for the presence of the free acid (apparently) only in these samples is not
known. This must represent one of the very few techniques for detection of
free sulfuric acid aerosol. Washing with water revealed the typical broad
"dust" band at 10 micrometers, particularly in Figure 35. This could account
for some of the wavelength shift. Returning to the sulfuric acid reference
spectrum, (Figure 34), traces 2 and 3 show the spectrum after exposure of
the sample to ammonia gas. The three broad bands are replaced by a single
band at 9.6 micrometers which is shifted to slightly longer wavelengths than
in Figure 33. The ammonium band is also shifted to slightly longer wave-
lengths.
Most spectra, and scores have now been recorded using the microreflector
technique, show a sulfate band resembling the salt rather than the free acid.
47
-------
o
H
rl
C <4-l
O
T3
C B
nj 3
W
O 4-1
H O
J-l 0)
3 &
"4-1 CO
H
3 4-J
CO Cl
QJ
,fl -H
4-1 ,0
H B
S Cfl
ni <1)
H 42
ti 4J
O
IS
(1) 0)
N CO
H CU
CO M
CU
42 OJ
4-1 CO
a o
CO O
VD
^H CU CM
O X
CO 4-1 T3
O C
VJ CU cfl
01 4-1
ctf O LO
0) CO
4-> CU
O CO >-(
Cfl T3 3
CX -H bO
B O -rl
H CO PI-I
in
01
CU
M
M
1-1
48
-------
cu
Q)
Cs
4-1
01
,0
CO
"B
rt
,0
0)
0)
0)
4-1
O
o
(fl
o
H
O
n)
CO
M
a)
3 6
w o
0) -H
*J S
U
n) CN
O. iH
e i
M OO
60
H
fn
49
-------
13
H !-i
CJ 0)
cfl 4-1
cfl
S-i M
3 (1)
LW 4J
rH 14H
3 cfl
en
o'S
4-1 Cfl
,0
H CO
0)
Cfl
H
6
H
CO
OJ O
4J J3
O CO
&
CO
0)
rH
co a
3 -H
& 4-1
e n
cfl cd
O &,
n
3 cfl
rH
C
o a)
co
a QJ
CU J3
4
-------
>,
iH
01
>
H
4-)
«
4-1
a
0)
4J
CO
t3
C
CO
o
O
CO
CO
H
H
CO T3
D -rl
G. O
6 nJ
tt)
O O
H
U 4-1
O H
3
C en
o
co
C a
CD
A! T3
nj 0)
0) -t-l
i-H 4-1
cx c
E oj
51
-------
SOUTHERN CALIFORNIA SURVEY
Ambient air samples were taken during a full day (about 6 A.M. to mid-
night) on a two-hour cycle at a number of locations around Southern California.
This produced well over 100 spectra, which are not easy to summarize since
all the nuances of the spectra cannot at this time be interpreted. Since
each location was sampled on only one or two days, these cannot be regarded
as representative of the various sites. While many spectra were similar to
those obtained at UCR, there were definite differences. The following list
gives the site and a brief commentary on each group of spectra.
Westwood UCLA - 26 June 1973;' 28 June 1973. Bands like sulfuric acid.
No evidence of nitrate.
Irvine UCI - 14 August 1973; 16 August 1973. Large particles: dust
band. Small particles: weak spectra.
Orange Co. APCD - 12 September 1973; 13 September 1973. Strong spectra.
Large particles: dust bands, nitrate bands.
Cerritos - 25 September 1973; 27 September 1973. Strong spectra on
25th, weak on 27th. Spectra on 25th similar to Riverside.
Pasadena Cal Tech - 17 July 1973; 19 July 1973. Strong dust bands,
nitrate bands, sulfuric acid, small CH band.
Claremont - 10 July 1973; 12 July 1973. Strong spectra, similar to
Riverside.
Green River (Santa Ana Canyon) - 19, 21 September 1973. Unusual spectra,
some without nitrate, some like sulfuric acid.
Palm Springs -7,9 August 1973. Mostly weak spectra; one clean nitrate,
sulfate spectrum.
Rim of the World (San Bernardino Mountains) - 24, 26 July 1973. Weak
spectra: sulfate bands.
Only a very few of the samples showed any indication of a CH band indicative
of organic material. To verify that the procedures used would show such a
CH band if such an aerosol were present, synthetic ozone-olefin aerosol was
generated and mixed with the ambient aerosol. A stream of 1000 ppm ozone at
100 ml/min was mixed with 50 ml/min of 3-ethyl-2-methyl-2-pentene in the
vapor phase to generate an aerosol which was mixed with ambient aerosol; then
52
-------
impacted and analyzed in the standard way. The spectra before water wash
(Figure 37) resembled the usual ambient air sample with no distinguishable
CH band. After water washing a distinct CH band as well as a variety of
longer wavelength bands appeared. Thus the method is capable of revealing
organic aerosol when there is enough present. This may be compared with the
spectrum of ambient aerosol taken on a day of very heavy haze with 12 million
particles per cubic foot; about as high a count as ever seen in Riverside
(Figure 18). This strong spectrum, which included distinctive nitrate bands
at 12 and 14 micrometers, was completely washed away by water. The strong
water vapor bands at 2.7 and 5.3 to 7.7 micrometers in the blank reflect the
high humidity of the ambient air that day.
EXTRACTION OF ORGANIC MATERIALS
Although the impaction spectra did not, in most cases, show evidence
of organic matter, it is evident that some organic matter is present never-
theless. To obtain spectra of the organic material, aerosol samples were
taken on high volume filters which were then extracted with various solvents.
To establish a relationship with pollution sources, reference samples were
taken of exhaust gases from various engines. These were treated with the
same solvents as the ambient air samples to obtain infrared spectra.
Equipment
The ambient air aerosol was collected with a Staplex SH810 hi-vol
sampler using a 8" Gelman-type A glass filter with a flow rate of 35 CFM.
The following modifications were made for the collections of auto exhaust
samples: A cone was attached to the hi-vol sampler with the point attached
to a "T" made of pvc tubing. One of the ports of "T" was connected to the
exhaust tail pipe. The third port was used as a by-pass.
Impaction samples of auto exhaust were taken by connecting the impactor
to a pvc "x". The second port was connected to the auto exhaust tail pipe.
The third port was connected to a cylinder of compressed nitrogen. The
nitrogen was used to cool the exhaust gases and prevent the melting of the
impactor wax seal and prevent water from condensing on the impaction plate.
The fourth port was used as a by-pass. The samples were collected on KRS-5
plates and scanned by the usual procedure.
53
-------
d
H
4-i
0)
.H O
O 4-1
Q) rH
d O
O 03
N O
O M
a)
a cd
o
H .H
4-) cd
cd s-i
^ 3
4-1 4-1
d cd
Q) C
a
G a)
o js
O 4-1
t>o a)
H >
42 O
4-1 S-l
H T3
S bO d
d ca
T3 -H JD
0) ,£3
X W iH
H cfl >,
B is d
o
iH J-l rQ
O 0) l-i
en 4J nj
O c« O
cd
CO
4-1
d 4J
OJ O
H a
d
cfl
o cd
M 0)
14-1 O
o d
o cd
0) -H
r-l 4J 0)
d, o >
S cd cd
cd a) CD
CO
bO
H
Pn
54
-------
Techniques
The hi-vol filter papers were extracted by placing % of a filter paper
into the soxhlet. 200 ml of solvent were placed in % liter flask and
refluxed for two hours. The solvent was then distilled off with the last
few mis being saved. The solution was taken up in an eye dropper and smeared
onto a KRS-5 plate. After the solvent evaporated, the sample was scanned on
the IR. The sample was first scanned in an environment of 100% humidity air,
and then dry air (0.5% humidity). Additional scans were run in an environment
of ambient air. First the sample was scanned, then washed with benzene,
scanned again and washed with water and then scanned again. The blank was
prepared by polishing the plate with tin oxide and methanol on a soft cloth.
Sampling
Twenty-minute hi-vol samples were taken from the exhaust of a 1970 gaso-
line powered Chevrolet pick-up (UCR vehicle 5-830). The truck was standing
and running at idle speed. The filter papers were cut in half and extracted.
A ten-minute impaction sample was collected.
Ten-minute hi-vol samples were taken from the exhaust of a 1960 Ford
5000 diesel tractor. The tractor was standing and running at idling speed.
The filter papers were cut in two and extracted. A five-minute impaction
sample was selected.
Twenty-minute hi-vol samples were taken up from the exhaust of a propane
powered pick-up (UCR vehicle 5-828). The truck was standing and running
at idle speed. The filters were cut in two and extracted. A 50-minute
impaction sample was collected.
These spectra are summarized in Table 4.
The polar solvents (water, methanol, acetone) produced spectra of ambient
aerosols similar to direct impaction with only marginal evidence for organic
molecules. The nonpolar solvents showed distinctive CH bands and bands in
the carbonyl stretch region. The various engine exhaust samples resembled
the ambient air spectra in a general way, but the agreement was far from
complete. Methods like these hold promise for distinguishing organic matter
emitted directly from engines, from that formed photochemically in the atmo-
sphere.
55
-------
TABLE 4. INFRARED BANDS (MICROMETERS) OF AMBIENT AND EXHAUST AEROSOLS
CH2C12 extraction
Ambient air
Gasoline exhaust
Diesel exhaust
Propane exhaust
CH3COCH3 extraction
Ambient air
Gasoline exhaust
Diesel exhaust
Propane exhaust
3.0
3.0
3.0
3.3
3.4
3.4
3.4
3.
3.
3.
3.
5.
5.
5.
5.
5
5
5
5
9
9
9
9
5.9
5.9
5.9
5.9
6.2
6.2
6.2
6.2
6.2
6.2
7.3
7.3
7
7
7
.0
.0
.6
7
7
7
7
8
.4
.4
.4
.4
.5
7.9 1.0-10.0
8.0 9.0-10.0
7.9
7.9
8.7 9.6-10.6
11.7
12.1
CH3CH2OCH2CH3 extraction
Ambient air
Gasoline exhaust
Diesel exhaust
CH3OH extraction
Ambient air
Gasoline exhaust
Diesel exhaust
Propane exhaust
C5H6 extraction
Ambient air
Gasoline exhaust
Diesel exhaust
H20 extraction
Ambient air
Gasoline exhaust
Propane exhaust
3.0
3.0
2.8
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
4
4
4
4
4
4
4
4
4
4
5.9
5.9
5.9
4.1
5.8
5.8
5.8
5.8
5.8
5.8
6.2
6.2
5.8-6.
6.3
6.3
6.2
6.2
6.2
6.3
6.3
6
6
6
3 6
6
6
6
6
6
6
.9
.9
.9
.8 -
.8
.8
.8 -
.9
.9
.9
7
7
7
7
7
7
7
7
.3
.3
.3
.8
.4
.4
.4
.4
7.8
9.6 12.0
9.6
9.6 12.0
7.9 9.0-10.0
9.0-10.0
8.0-10.0
8.0-10.0
8.0-10.0
11.7
11.7
1.3 mm slit impaction
Ambient air
Gasoline exhaust
3.0 -
3.
2
6.2
7.0-8.0
7
.0
12.7
0.38 mm slit impaction
Ambient air
Gasoline exhaust
3.0-3
.2
5.9
5.8
6.2
6.3
7
.1
7
.6
9.3
9.4
12.0
Diesel exhaust
3.3
12.0
56
-------
In addition to these spectral studies, some carbon/hydrogen analyses
have been obtained by the classical combustion technique. For this analysis
glass fiber filter samples were obtained over a 96-hour period from May 11-
15, 1973. Each of six filter strips was weighed before and after sampling
and again after soxhlet extraction with one of five solvents. The extracts
were reduced in volume by evaporation and 15 ml of 50 ml concentrate was
evaportated onto an infrared window to obtain an infrared spectrum. The
remaining 35 ml of extract were evaportated on pre-weighed glass fiber filters.
After all the solvent had evaporated, reweighing gave the weight of soluble
material deposited on the filter. These were then sent out for combustion
analysis. A blank was run on the filter paper. Table 5 gives data for these
experiments. The principal conclusion is that the proportion of organic
TABLE 5. CARBON/HYDROGEN ANALYSIS OF AMBIENT AEROSOL EXTRACTS
Solvent
Ether
McCl2
C6H6
Acetone
MeOH
Weight*
mg
198
130
134
70
155
Wt. loss
extraction
mg %
41
33
24
64
105
20.
25.
17.
91.
67.
6
4
9
4
7
Extract
%C %H
13.
12.
23.
0.
2.
4
9
4
3
2
7.
4.
4.
0.
2.
9
4
4
5
1
% of total
sample
%C %H
2
3
4
0
1
.8
.3
.2
.3
.5
1.
1.
0.
0.
1.
6
1
8
5
4
* q
In approximately 940 m3 of air sample
carbon is quite small. The hydrogen analysis is much more uncertain since
there was no way to insure that all the water had been removed from the paper
prior to analysis. The C/H atom ratios are not very consistent or easily
interpreted in terms of organic structures. Since visual examination of the
aerosols always shows black particles which are probably mostly elemental
carbon, an attempt was made to extract all the organic carbon by extracting
a filter, first with ether, then with methylene dichloride, then with benzene
and finally with methanol. These samples contained 22.5% carbon and 6%
57
-------
hydrogen before extraction and 26.7% carbon and 5.9% hydrogen after extrac-
tion. Another 24-hour hi-vol sample contained 11.4% carbon and 7.4% hydro-
gen. These analyses are consistent with the view that there is considerable
elemental carbon in the ambient aerosol.
PHOTOCHEMICAL FORMATION OF AEROSOLS
A complete account of photochemical aerosol formation would include a
demonstration that realistic aerosol concentrations can be formed from
realistic concentrations of reactants. This has not been clearly demonstrated
for any of the known aerosol-forming mechanisms. Several laboratories have
shown aerosol formation at high (ppm) concentrations of olefins irradiated
with nitrogen oxides. In our earlier work spectra of such aerosols were
obtained which showed clear CH bands. To extend this work to lower concen-
trations a series of experiments were done with the flow tube photoreactor
using a-pinene and N02 as reactants. Concentrations down to 1/32 ppm (31 ppb)
of a-pinene were used with residence times up to eight hours and a 2/1 ratio
(by moles) of the hydrocarbon to N02. The aerosol was counted with the
Bausch and Lomb 40-la dust counter for particles larger than 0.3 micrometer
in diameter.
Compressed air from the house supply was purified with a heatless "dryer"
in which charcoal was used in place of the desicant. After leaving the
purifier this air was humidified by passage through an ice bath. It was found
necessary to pass the air through a 0.3 micrometer filter to obtain a zero
background count on the air entering the reactor. It was also important to
prevent back flow and diffusion of room air into the reactor outlet. Counts
made under these conditions showed no countable particles entering the
reactor from the input manifold and showed the same reading at the reactor
exit (dark). The reactants were dispensed into the airstream from 35-liter
low pressure breathing oxygen tanks. Although the reactants were made up by
well established procedures, that is adding liquid to a "low pressure oxygen"
(LPO) tank and pressurizing it with nitrogen, and the input concentrations
are calculated by the ratio of input reactant stream to input airstream, no
actual measurements of the reactants were made and therefore there is some
uncertainty as to accuracy of the concentrations. The results are summarized
58
-------
in Table 6.
TABLE 6. AEROSOL PRODUCED BY THE PHOTOCHEMICAL REACTION OF NITROGEN DIOXIDE
AND a-PINENE
Residence
time
(hours)
1
1
1
1
1
1
1
2
2
2
2
2
2
4
4
4
4
4
8
8
8
NO 2
cone.
(ppm)
1/2
1/4
1/4
1/8
1/8
1/16
1/32
1/4
1/4
1/8
1/16
1/32
1/16
1/4
1/8
1/8
1/16
1/32
1/16
1/32
1/64
a-Pinene Count (million particles/ft.3)
cone. Diam.
(ppm) (micron) 0.3 0.5 1.0 1.8
1
1/2
1/2
1/4
1/4
1/8
1/16
1/2
1/2
1/4
1/8
1/16
1/32
1/2
1/4
1/4
1/4
1/16
1/8
1/16
1/32
2.492
.2603
.4865
.0291
.0085
.0041
.0000
1.355
.9697
.0877
.0807
.0096
.0000
5.317
2.077
.5088
.0838
.0012
.0827
.0523
.0011
.1903
.0042
.0121
.0000
.0000
.0000
.0443
.0310
.0000
.0000
.0001
.7324
.2238
.0076
.0013
.0001
.0028
.0025
.0000
.0077
.0000
.0000
.0008
.0000
.0704
.0086
.0086
.0000
.0000
.0005
.0004
.0000
.0000
.0000
.0000
.0000
.0000
Reproducibility was only fair. Aerosol counts varied by a factor of two to
four between duplicate runs.
Table 6 shows that aerosol can be formed in this system but only in large
count at unrealistically high concentrations. Longer irradiation times pro-
mote greater aerosol formation but even for eight hours time the amount of
aerosol formed was tiny at the lowest concentrations. It was apparent from
these experiments that a realistic simulation of smog aerosol generation
would require a far more sophisticated laboratory reactor design.
59
-------
Additional experiments were done to test the idea that N03 might be able
to oxidize S02 to S03:
N02 + 03 -> N03 + 02
N03 + S02 -» S03 + N02
S03 + H20 -> H2SO(+ (aqueous aerosol)
The sulfuric acid should form as a countable aerosol. The reagents were
mixed and given time to react, then the number of aerosol particles in the
mixture was counted.
The stirred flow reactor (Figure 7) was used as the reaction vessel.
The air that was used to dilute the reagents was purified passing through a
heatless air dryer filled with charcoal. The air was then brought up to
constant humidity by bubbling through distilled water immersed in an ice
bath.
After a five-hour flush of the vessel with air, an aerosol count was
.taken with a 40-1 Bausch and Lomb counter to determine the background concen-
tration of aerosol. This background varied from (0.0 to 0.2) x 1Q3 particles/
cu ft. The background was not altered when N02 or 03 were introduced into
the vessel. Introduction of S02 did alter the background, so a filter was
connected to the line that transfers S02 from its dispensing tank to the
input manifold. All the reagents were stored in LPO tanks in which they were
diluted with prepurified nitrogen.
At the start of the experiment the aerosol count was taken, then the
reagents were mixed with an appropriate amount of air to give the desired
concentrations. Nitrogen dioxide was measured colorimetrically using a spec-
tronic 20 and Salzmann's reagent. Ozone was measured with a Mast Ozone meter
and sulfur dioxide with a conductivity analyzer. Because of mutual inter-
ferences in the analytical technique the individual reactants were added
to a stream of clean air and their concentrations measured, then the three
components were dispensed into the manifold.
Two experiments were run with a concentration of 1 pptn of all regents
and 10 ppm of all reagents. Both times reactor output did not exceed the
60
-------
background particle count. It can be concluded that S(>2 does not undergo
oxidation to S03 under the condition of this experiment. Calvert and McQuigg
(12)
have reached the same conclusion.v '
With this stirred flow system and purified air the PAN/S02 was reinvesti-
gated. As before, aerosol could be produced by photolysis of this mixture
but not in the dark. Neither this reaction nor the photodecomposition of PAN
is understood. The absorption of near UV (about 3000) by PAN is quite weak
and seems inadequate to account for the photolysis. This cannot be a firm
conclusion until quantitative data become available on the strength of this
weak UV absorption. The PAN/SC>2 experiment does demonstrate that this experi-
mental system can produce an aerosol.
61
-------
SECTION 4
FUTURE POTENTIAL
The techniques evolved under this program have good potential for further
useful development. The infrared technique especially, if further developed,
might be most useful. Its chief advantage is the possibility of obtaining
unique information on such difficult aerosol components as sulfuric acid
with the potential for distinguishing it from sulfates. Aside from a direct
gas phase technique such as the spectrophone (not yet perfected) impaction
onto a noble metal such as platinum offers less risk of sample transformation
than any other technique. Its principal shortcoming is the lack of a quanti-
tative calibration. The following developments might be investigated to
improve the method:
1. Quantitative calibration of the infrared spectrum by sampling from
a synthetic aerosol stream analyzed by some other method or by
microanalysis of the impacted spot.
2. Improvement of the air environment during spectral recording by
better C02 and water removal.
3. Improvement of the library of reference spectra by generation of
better synthetic aerosols by either a) mixing acid gases such as
sulfuric, acetic, etc. with basic gases such as ammonia or organic
amines. This would permit inclusion of organic components b) photo-
chemical aerosols generated in a smog chamber such as one of the
two at UCR.
The impactor itself might be improved several ways:
a) Shorten the impaction slit to reduce the length of the sample
to match the image in the reflector attachment. This would -
be especially valuable for sampling photochemical synthetic
aerosols where sample volume is limited.
62
-------
b) Impact on a small drum which could be turned stepwise every
hour (as in the modification of the Lundgren impactor made by
Cunningham at Argonne).
c) Better sample flow control and measurement using either a
critical oriface or a surge tank or both.
Two "apparent" improvements that would probably not be worthwhile are:
a) Adding a third impaction stage for further size discrimiation.
This is based on the belief that particles large enough to
impact have a bimodal distribution. The photochemical aerosol
which is responsible for visibility loss is primarily in the
diameter range below one or two micrometers and the soil dust
is primarily in the larger size fraction.
b) Increase in resolution of the spectrum. Tests at EPA with
Fourier interferometry indicate that there is no additional
structure to be resolved. This is to be expected for these
condensed phase spectra.
63
-------
REFERENCES
(1) Stephens, E. R. and M. A. Price. Smog Aerosol: Infrared Spectra.
Science, 168 (3939), 1584 (June 26, 1970).
(2) Stephens, E. R. and M. A. Price. Comparison of Synthetic and Smog
Aerosols. J. Colloid Interfac. Sci., (1972) 1 (4) 39, 272 Aerosols
and Atmospheric Chemistry, Acad. Press. New York, 1972, Ed., G. M.
Hidy, p. 167.
(3) Haagen-Smit, A. J. Engineering and Science Dec, 1950 (a Cal Tech
magazine), Bay Area APCD microfilm #448. McCabe, L. C. Ind. and
Eng. Chem., v. 43, No. 4, April 1951, p. 89A.
(4) Martens, A. E. and J. D. Keller. J. Amer. Indust. Hygiene Assoc.,
v. 29, p. 257 (1968).
(5) Cunningham, P. T., S. P. Johnson and R. T. Yank. Variations in
Chemistry of Airborne Particulate Material with Particle Size and Time.
Environ. Sci. Technol., 8 (2), February 1974, p. 131.
(6) Cunningham, P. T. and S. A. Johnson. Spectroscopic Observation of Acid
Sulfate in Atmospheric Particulate Samples. Science, v. 191, January
1976, p. 77.
(7) Stephens, E. R. and M. A. Price. Atmospheric Photochemical Reactions
in a Tube Flow Reactor. Atmos. Environ., 3 (573) (1969).
(8) Scott, W. E., E. R. Stephens, P. L. Hanst and R. C. Doerr. Further
Developments in the Chemistry of the Atmosphere. Proc. API 37 III,
171-183 (1957).
(9) Varetti, E. L. and G. C. Pimentel. Spectrochem. Acta, 30A, pp. 1069-
1072, 1974.
(10) Lachowicz, D. R. and K. L. Krenz. J. Org. Chem., 32, 3885 1967.
(11) Stephens, E. R. Hydrocarbon in Polluted Air. Coordinating Research
Council, Project CAPA 5-68, Summary Report, June 1973. NTIS Accession
No. PB 230, 993/AS.
(12) Calvert, J. G. and R. D. McQuigg. Int. J. Chem. Kinetics Symposium #1,
1971, p. 146 (1975).
64
-------
BIBLIOGRAPHY
Altshuller, A. P. Regional Transport and Transformation of Sulfur Dioxide
to Sulfate in the U.S. J. Air Pollut. Control Assoc., 26 (4), April
1976, p. 318.
Appel, B. R., P. Colodny and J. J. Weselowski. Analysis of Carbonaceous
Materials in Southern California Atmospheric Aerosols. Environ. Sci.
Technol., 10 (4), April 1976, p. 359.
Flocchini, R. G., T. A. Cahill, 0. J. Shadoan, S. J. Lange, R. A. Eldred,
P. J. Fenney, G. W. Wolfe, D. C. Simmerath and J. K. Suder. Monitoring
California's Aerosols by Size and Elemental Composition. Environ. Sci.
Technol., 10 (1), January 1976, p. 76.
Grosjean, D. M. and S. K. Friedlandler. Gas Particle Distribution Factors
for Organic and Other Pollutants in the Los Angeles Atmosphere. J. Air
Pollut. Control Assoc., 25 (10), October 1975, p. 1038.
Husar, R. B., W. H. White and D. L. Blumenthal. Direct Evidence of Hetero-
geneous Aerosol Formation in Los Angeles Smog. Environ. Sci. Technol.,
10 (5), May 1976, p. 490.
Levaggi, D. A., J. S. Sandberg, M. Feldstein and S. Twiss. Total Anthropo-
genic Suspended Particulate as Derived from Chemical Analysis of Chloride
and Silicate on High Volume Samples. J. Air Pollut. Control Assoc.,
26 (6), June 1976, p. 554.
Maddalone, R. F., R. L. Thomas and P. W. West. Measurement of Sulfuric Acid
Aerosol and Total Sulfate Content of Ambient Air. Environ. Sci. Technol.,
10 (2), February 1976, p. 162.
Miller, D. F. and A. Levy. Exhaust Hydrocarbon Relationships with Photo-
chemical Aerosol Formation. J. Air Pollut. Control Assoc., 26 (8),
August 1976, p. 778.
Miller, D. F., A. Levy, D. Y. H. Pui, K. T. Whitby and W. S. Wilson, Jr.
Combustion and Photochemical Aerosols Attributable to Automobiles.
J. Air Pollut. Control Assoc., 26 (6), June 1976, p. 576.
65
-------
BIBLIOGRAPHY (cont.)
O'Brien, R. J., J. R. Holmes and A. H. Bockian. Formation of Photochemical
Aerosol from Hydrocarbons,. Chemical Reactivity and Products. Environ.
S._i. L'eclmol., 9 (6) June 1975, p. 569.
O'Brien, R. J., J. H. Crabtree, J. H. Holmes, M. C. Hoggan and A. H. Bockian.
Formation of Photochemical Aerosol from Hydrocarbon Atmospheric Analysis.
Environ. Sci. Technol., 9 (6) June 1975, p. 577.
Roberts, P. T. and S. C. Friedlander. Photochemical Aerosol Formation S02,
1-Heptene and NOX in Ambient Air. Environ. Sci. Technol., 10 (6), June
1976, p. 573.
Sandberg, J. S., D. A. Levaggi, R. E. DeMandel and W. Siu. Sulfate and
Nitrate Particulates as Related to S02 and NOX Gases and Emissions.
J. Air Pollut. Control Assoc., 26 (6), June 1976, p. 559.
Tanner, R. L. and L. Newman. The Analysis of Airborne Sulfate: A Critical
Review. J. Air Pollut. Control Assoc., 26 (8), August 1976, p. 737.
White, W. H. and R. B. Husar. A Lagrangian Model of the Los Angeles Smog
Aerosol. J. Air Pollut. Control Assoc., 26 (1), June 1976, p. 32.
66
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 tW-W/3-77-044
2.
4. TITLE AND SUBTITLE
FORMATION OF PHOTOCHEMICAL AEROSOLS
6. PERFORMING ORGANIZATION CODE
3. RECIPIENT'S ACCESSI OF> NO.
5. REPORT DATE
May 1977
7 AUTHOR(S)
Edgar R. Stephens and Monty A. Price
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Statewide Air Pollution Research Center
University of California
Riverside, California 92502
10. PROGRAM ELEMENT NO.
1AA603 AJ-01 (FY-76)
11 CONTRACT/GRANT NO
Grant No. 800868
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory-RTF, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final .
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The objective was to develop a better understanding of smog aerosol formation
with particular reference to haze in the Southern California area. This study
combined laboratory work with ambient air studies. Counting of particles by light
scattering was the principle physical technique while infrared analyses was the
major source of chemical information. A new reflectance spectroscopy technique
was also developed.
Infrared spectra of ambient aerosols have bands assigned to sulfate, nitrate,
ammonium, and water which are completely removed by water washing of the sample
but not by benzene. Synthetic aerosols generated by mixing ammonia with sulfuric
and nitric acids produce similar spectra.
Many ambient aerosol particles are hygroscopic or deliquescent so that they
swell as the relative humidity increases and shrink as it decreases. This readily
explains the heavy haze seen at high humidity days.
A variety of reaction products (SO ~
the requirement of water solubility and hygroscopicity to contribute to haze and loss
of visibility. It is concluded that direct control of aerosol emissions will not
markedly improve visibility in Southern California.
NO,, organics) can potentially fulfill
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Air pollution *
Aerosols
Smog
Photochemical reactions
Tests
Infrared spectroscopy
Humidity
Field tests
Southern California
13B
07D
04B
07E
14B
13. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
77
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
67
-------
|