EPA-650/4-74-009
DECEMBER 1973
Environmental Monitoring Series
-------
EPA-650/4-74-009
ENVIRONMENTAL CHAMBER STUDIES
OF ATMOSPHERIC AEROSOLS
by
D.F. Miller and Arthur Levy
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
Contract No. 68-02-0574
Project No. 21 AKB-9
Program Element No. 1AA008
EPA Project Officer: Dr. W.E. Wilson
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
December 1973
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
1 1 .
-------
ACKNOWLEDGMENTS
The aLithors gratefully acknowledge D.Y.H. Pui, Dr. K. T. Whitby, and Dr. B. K. Cantrell at
the University of Minnesota for their assistance in measuring aerosol size distributions.
The authors also wish to acknowledge their fellow workers who made contributions to this
program, especially D. W. Joscph, G. W. Keigley, D. A. Trayser, and G. F. Ward.
Finally, we wish to acknowledge the leadership and interest of Dr. W. E. Wilson, Jr., EPA
Project Officer.
ii’
-------
TABLE OF CONTENTS
Page
INTRODUCTION AND OBJECTIVE
SUMMARY OF OBSERVATIONS AND CONCLUSIONS.
EXPERIMENTAL
RESULTS
3
. 8
10
Aerosol Formation and Eye Irritation
Mechanisms of Aerosol Formation in Smog
Physical Processes
Chemical Processes
Influence of Primary Auto-Exhaust Aerosols on Development of
Secondary Aerosols
REFERENCES .
Table I. Summary of Results of Smog-Chamber Experiments. . 9
Table 2. Correlations Between Eye irritation and Various
Independent Variables
Table B-I. Summary of Aerosol Collections by Filtration
3
Presentation of Data
Summaries of Data
DiSCUSSION
10
10
II
12
17
20
A-i
B-I
APPENDIXES
A. EXPERIMENTAL METHODS
B. SUMMARY OF AEROSOL COLLECTIONS. .
LiST OF TABLES
11
B-i
V
-------
FINAL REPORT
on
ENVIRONMENTAL CHAMBER STUDIES OF
ATMOSPHERIC AEROSOLS
to
U.S. ENVIRONMENTAL PROTECTION AGENCY
from
BATTELLE
Columbus Laboratones
December 12, 1973
• INTRODUCTION AND OBJECTIVE
l’he objective of this research program is to obtain data that can be used by EPA to
evaluate the role of pnmary auto-exhaust aerosol in the subsequent formation of photochemical
aerosol and to determine what relationships if any might exist between photochemical-aerosol
formation and eye irritation. In pursuing the objective, experiments were conducted in which
synthetic hydrocarbon-NOn-air mixtures and automobile exhaust-air mixtures were irradiated in a
61 0-cu-ft smog chamber. Emphasis was placed on strict control of all experimental parameters,
on complete detailed analyses of the formation of photochemical aerosols* and on the chemistry
associated with the formation of other photochemical-smog manifestations, including eye irrita-
tion. In addition to obtaining data pertinent to the stated objectives, data were obtained which
revealed chemical mechanisms important in aerosol formation and physical mechanisms account-
ing for losses of aerosols in smog chambers.
SUMMARY OF OBSERVATIONS AND CONCLUSIONS
In addition to this report, three papers have been presented which describe various aspects
of the program. The papers are entitled, “Aerosol Formation in Photochemical Smog. The Effect
of Humidity and Small Particles” 1 , “Evolution of the Freeway Aerosol” 2 , and “Effect of
Stirring on Aerosol Production in a Smog Chamber” 3 A fourth paper entitled “The Size
Distribution and Concentration of Combustion and Photochemical Aerosols Attributable to
Automobiles” t4 will be presented at the 67th Annual Mc eting of the Air Pollution Control
Association, June, 1974.
t in this report, the term photochenucal ,ierosois refers to aerosols (condensed n atter) which form as a resuit of photochemicjiiy
induced reactions occurring in the smog chamber
-------
2
Although these papers discuss many of the results of this program, several other important
observations were made while assembling and perusing these data. The observations are listed
below, together with some of the conclusions contained in the papers cited. Observations nol
referenced are discussed in the text of this report.
• There is no substantial relationship between the severity of eye irritation and the
amounts of aerosols formed photochemically.
• Inclusion of primary aerosols during the formation of photochemical aerosols
changes the size distribution of the resulting aerosol and may have a sizable
influence on light scattering and visibility reduction. However, the presence of
primary aerosols appears to have little effect on the volume of aerosol material
produced by thc photochemical processes.
• Losses of aerosols to chamber surfaces agree with diffusion-rate theory for various-
size particles However, turbulence created by mechanically stirring the chamber’s
atmosphere increases the loss rate by factors as large as 5, and the loss rates differ
for aerosols of different chemical composit ons. 131
• The relationships between aerosol mass concentration inferr d from size-distribution
data, aerosol mass concentration determined gravimetrically, and aerosol mass con-
centration predicted by total light scattering are fairly consistent and rarely differ by
factors greater than
• Patterns of photochemical-aerosol formation from aromatic hydrocarbon-NO -air
and automobile exhaust-air mixtures are similar, and OH radicals appear to be
important precursor intermediates for aerosol formation in these cases. With olefinic
hydrocarbon- NO -air and olefinic hy drocarbon-NO -SO 2 -air mixtures, photoche mical-
aerosol formation appears to depend on ozone concentrations.
• The initial concentration of aerosols emitted by automobiles is very high and the
aerosol sizes are quite small ( O.02 pmdiamcter). Under most laboratory conditions.
the small aerosols coagulate quite rapidly unless rapid and extensive dilution is
applied. 141
• TEL motor-fuel additive has no pronounced effect on the rate of NO photo-
oxidation or on other photochemical-smog manifestations, including eye irritation
Presumably, additional important information can be derived from the data furnished in this
report. Chemical models of the smog-chamber data or chemical analyses of the numerous aerosol
collections may, for example, lead to an improved understanding of aerosol development in
smog.
EXPERIMENTAL
Battelle-Columbus’ 610-cu-ft smog chamber has been described previously Light-intensity
measurements by NO 2 photolysis 6 and o-nitrobenzaldehyde photolysis 17 1 agree quite well, as
described by Gordon. 181 The value for kd is—’0.45 min 1 . Total hydrocarbon was determined by
-------
3
flame ionization, specific hydrocarbons by flame-ionization gas chromatography, (‘0 by NDIR.
SO 2 by flame photometry and a coulomctric analyzer, 03 by chemiluminescenec with ethylene.
NO and NO 2 by automated Saltzman using a dichromate oxidizer for NO, PAN by electron-
capture gas chromatography, and light scattering by broad-band (420 to 570 nm) integrdted
nephelometry. Measures of eye irntation were made by three 7-member eye-panel teams chosen
from Battelle-Columbus’ staff. Selection and screening of the panelists are described in
Appendix A.
Aerosol size distributions were obtained using the Minnesota Aerosol Analyzing System
(MAAS) developed by Dr. K. T. Whitby and his colleagues at the University of Minnesota.
Instruments comprising the MAAS are a portable electrical-mobility analyzer, a modified optical
particle counter, and a condensation nuclei counter. Calibration and operation of this system in
determining size distributions has been reported by Whitby et al.t 9
Synthetic automobile exhaLists used as part of the total exhausts in Runs AA-032 through
AA-038 were made up of 50 percent by weight of the respective motor fuel (simulating
fuel-derived exhaust constituents) and 50 percent by weight of a C 2 -C 4 hydrocarbon mixture
(simulating combustion-derived exhaust constitutents). The percentage composition of the coin-
bustion mixture was ethane 4.0, ethylene 44.1, acetylene 28.0, propylene 16.0, l-butene 6.2,
and trans-2-butene 1.7
Procedures for operating the automobiles and transferring exhausts to the smog chamber are
described in Appendix A.
RESULTS
Presentation of Data
A major requirement of this program was to furnish EPA a comprehensive copy of data
emanating from the aerosol and chemical measurements made during smog-chamber experiments
Because of the large amount of data obtained, a separate Data Report was prepared containing
(1) Time-concentration profiles of the gas-phase chemistry of each smog experiment
(2) Computer tabulated printouts of experimental conditions and rates of change of
various constituents during each experiment
(3) Time-concentration profiles of various aerosol parameters inferred from aerosol size
distributions measured during the experiments
(4) Computer-generated plots of the changes in aerosol surface distributions with
irradiation time
(5) Tabulations of the response times to eye irritation for each panelist, and the
arithmetic and geometric mean times for each experiment.
Two copies of the Data Report were delivered to EPA together with this report. Examples of the
gas-phase and aerosol profiles appearing in the Data Report are. shown in Figures 1 through 4.
-------
Ozone
0.400
ppm
- -
Carbon
monoxide
20.000
ppm
—— — ——
Sulfur
dioxide
0.020
ppm
———-
Nitric
oxide
l 000
ppm
——
Nitrogen dioxide
0.600
ppm
Run AA 032
NO 2
— -
/
/
/
9’
8’
60
4 ’
N
N
/
U )
C
.4-
0
4-
C
a ,
C .,
a,
0
P t’
P t’
5’
‘4
‘ 4
CO
NO
N
St .
‘4
2
N
‘ 4%
‘S
‘ 4%
a
a
a a a a — — —
0 25 50 75 100 125 150 175 200 225 250 275 300 325 350
Irradiation Tsme,min
FIGURE 1. SMOG PROFILE OF INORGANIC GASES
-------
50 75 100 125 150 175 200 225 250 275 300 325 350
Irradiation Time,min
Total hydrocarbon 20.000 ppm
Methane 2.000 ppm
PAN 0.080 ppm
FIGURE 2. SMOG PROFILE OF ORGANIC GASES
Run 44 032
— — — — — — — — — — — —
100
90
80
70
60
50
40
30
0
-9
U)
4-
C
0
4-
C
a,
C.)
a)
/
,
/
,
,
/
,
,
,
,
0 25
‘ I l
-------
V-Volume / Lm 3 /cc 30
S-Surface /.Lm 2 /cc 2000
N-CNC number(No./cc) 800000.
B-BSCAT (el.OE-4/m)0.500
T—Surface >.I/.L( 1 u.m 2 /cC) 700
FIGURE 3. SMOG PROFILE OF INTEGRATED LIGHT SCATTERING (BSCAT) AND VARIOUS AEROSOL
PARAMETERS COMPUTED FROM SIZE-DISTRIBUTION MEASUREMENTS
100
Run 44 032
U)
4-.
C
-I
0
4-
C
w
C-)
w
0
0 25 50 75 100 125 150 175 200 225 250 275 300 325
Time, mm
-------
5000 Run AA 032
Irradiation Time,min
A- 0.00
B- 22.00
4000- C- 52.00
D- 102.00
E- 182.00
F -272.00
U
3000-
I
- 2000-
1000 - /B/
0 _____ __
0.001 0.010 0.100 1.000 0.000
Particle Diameter (Dp)i/.Lm
FIGURE 4. CHANGES IN AEROSOL SURFACE DISTRIBUTION AS A FUNCTION OF IRRADIATION TIME
-------
8
Figures I through 4 describe results from an experiment where filtered auto exhaust
(nonleaded fuel) was irradiated for 6 hours. Just before irradiation, condensation nuclei were
added to the chamber. In Figure 1, the loss rate of SO 2 is shown to be less than the dilution
rate (indicated by the CO curve). This apparent anomaly reflects the inaccuracy of the SO 2
monitor at very low concentrations ( 0 02 ppm). In Figure 2, no loss of methane is indicated
because the purified air used to make up volume losses contains methane. In Figure 3, the
number concentration of aerosols is greatest at 0 minutes because of the addition of condensa-
tion nuclei (CN) prior to irradiation. Although the number concentration is high (7 x l0 cm- 3 )
at t = 0 minutes, the initial amount of aerosol surface due to CN is small (‘-50 g .im 2 cm- 3 ), as
indicated in Figures 3 and 4. In this case, the added CN had little effect on the development of
the photocheniically derived aerosols.
Aerosol Mass and Chemical Data. Numerous aerosol collections were made at the end of the
irradiation periods (4 to 6 hours). Aerosols were collected on a varicty of filter media for mass
determinations and/or subsequent chemical analyses. Data pertinent to these collections are
presented in Table B-I, Appendix B.
Computer Tape Record. In addition to the Data Report, a 7-track IBM compatible
magnetic tape was prepared containing all gas-phase data converted to engineering units. The tape
WdS delivered to the Project Officer, September 26, 1973.
EPA-CRC CAPE-19 Study. As an adjunct to this program, a number of aerosol analyses
using the MAAS were conducted to characterize the aging process of primary automobile
particulate in a dark residence chamber. Those analyses were performed in conjunction with the
EPA-CRC CAPE-19 program at BCL. Copies of those data, tabulated and graphically displayed
by computer, were delivered to the Project Officer, April 6, I 973
Summaries of Data
Initial experimental conditions and results of all the smog-chamber experiments are
summarized in Table 1. LeSS familiar abbreviations are defined below
% RH percent relative humidity
bscat extinction coefficient of light scattering
ST = total aerosol surface concentration
VT = total aerosol volume concentration
Mass = aerosol mass concentration determined by filtering and weighing
I AN = peroxyacetylnitrate
I-ICUO = formaldehyde
RCHO = total aldehyde.
-------
TABLE 1 SUMMARY OF RESULTS OF SMOG-CHAMBER EXpERIMENTS(aI
Maximum or Final Concentrations
Initial Concentrations and Conditions Eye
-— Aerosols Stirring Gases _______— Aerosols — Irritation as
H 2 0 CO. NO, NO 2 SO 2 03. bscat. ST. Speed. 03. PAN HCHO. RCHO b at ST. VT. Mass. Response
Run Hydrocarbons, ppm C % RH ppm ppm ppm ppm ppm 10 4 m 1 pm 2 cm rpm ppm ppm ppm ppm 10 4 m 1 pm 2 cm pm 3 cm pg m 3 Time sec
fai Abbreviations delined on pays 8
fbi Dark reactions
(c) Condensation nuclei added before irradiation
(di Auto exhaust filtered of primary aerosol
fe Leaded fuel, all other exhaust e’perirx nts made with unledded fuel
AA 005
Toluene, 466
66
<3 110 100
000 000
02
—
0
055 030
01
11
82
921
77
AA 006
Toluene. 46 3
58
<3 098 098
000 000
02
<10
0
04S 040
0 1
1 3
100
9,274
640
1.100
56
AA 007
Toluene, 469
66
<3 094 094
000 000
0 2
<10
990
053 043
0 1
1 4
21
1.566
205
278
47
AA 008
1 heptene, 358
66
<3 072 076
000 000
02
<10
0
064 005
1 3
48
13
7.123
352
180
46
AA.009
1 heplene, 203
66
<3 072 072
000 000
02
<10
0
064 007
06
28
17
2.196
93
67
84
AAO1O
1 heptene 346
70
<3 070 070
000 000
02
—
1.000
062 001
12
51
09
3.042
100
80
61
AAO11
Benzene, 308
70
<3 071 072
000 000
02
<10
0
000 —
00
03
29
2,316
102
117
189
AA 012
Toluene 16/mesitylene. 28
66
<3 068 069
000 000
02
<10
0
098 020
04
2 2
21
8,452
443
321
83
AA013
Benzene, 321
66
75 075 075
000 000
02
<10
0
016 —
00
02
21
2,013
89
—
199
AA015
Bensene, 318
62
<3 073 071
010 000
02
<10
0
000 —
00
00
58
4,019
198
140
198
AAO16
—
58
<3 <005 000
019 000
02
<10
0
006 —
—
—
65
6.372
234
174
—
AAO17
—
14
<3 <005 000
023 000
02
<10
0
000 —
—
—
17
2,904
117
—
218
AA O I8(b)
—
66
<3 020 200
060 060
02
<10
0
— —
—
—
02
1,478
28
—
—
AA 01 g(b)
Propylene, 90
70
<3 <006 000
096 072
02
<10
0
— —
—
—
12
17,580
518
263
—
AA 020
a pinene, 27 4
70
<3 073 076
000 000
02
<10
0
044 —
—
—
>100
>20.000
>2,000
—
97
AA021 ’ 1
I heptene, 200
78
<3 <005 —
000 064
02
<10
0
— —
—
—
28
7,553
248
65
—
AA 022 (b)
1 heptene, 208
78
<3 <005 —
0 5 051
02
<10
0
— —
—
—
20
>20.000
>600
281
188
AA 023 1c)
Toluene, 18 3
62
<3 076 073
000 000
02
100
0
042 025
0 1
09
20
5,592
316
456
92
AA-024
Auto exhaust. 160
55
225 200 048
006 000
—
1,292
0
052 025
05
18
6
3,353
157
165
88
AAO2S(d)
Auto exhaust, 160
58
245 140 089
001 000
02
<10
0
078 049
05
18
2
3.542
114
125
104
AA.028
1 tieptene, 120,
66
<3 043 001
006 000
02
<10
0
039 —
03
1 4
09
3,028
77
102
103
AA 029
1 heptene, 12 1
74
<3 0 48 001
006 000
02
<10
1.000
0 37 —
0 3
1 3
06
2,470
67
84
92
AA 030
1 heptene, 11 2
66
<3 052 001
000 000
02
<10
0
051 —
03
1 3
03
509
9
52
118
AA03I 1 ’
1 heptene 117
70
<3 050 001
000 000
03
100
0
039 —
03
14
03
461
12
62
92
AA 032 (c.d)
Auto exhaust, 80
62
IS 097 022
<001 000
02
50
0
026 007
03
10
04
1,030
24
58
126
AA033
Auto exhaust,8 1
62
40 072 006
<001 000
02
60
0
053 019
03
12
04
1116
25
58
104
AA034(d)
Auto exhaust,82
58
41 078 016
<001 000
02
<10
0
054 017
04
11
04
756
16
47
114
AAO35IeI
Autoeyhaust,81
66
39 073 002
<001 000
03
512
0
046 022
03
Il
05
877
27
43
107
AA 035 (eI
Auto exhaust, 8 1
58
40 079 001
<001 000
03
433
0
050 023
04
1 2
05
827
25
38
138
AA 037 (de)
Auto exhaust. 80
58
42 077 009
<001 000
02
<10
0
057 029
03
1 2
03
755
16
42
94
AA 038 (d,e)
Auto exhaust, 82
58
40 018 009
<001 000
02
<10
0
045 027
03
1 2
03
670
15
46
103
‘ .0
-------
10
DISCUSSION
Aerosol Formation and Eye Irritation
One of the objectives of this program was to determine whether a relationship exists
between the degree of eye irritation and the concentration of photochemically generated
aerosols. The procedure for determining eye irritation was to expose two groups of panelists
(totaling 14 persons) to the chamber contents about 2 hours after the NO 2 concentration passed
through its maximum. Typically, during this period of irradiation, aerosol surface and volume
concentrations also reached or passed through their maximum values. It has been established on
the EF-8 project for APi that this period of irradiation coincides with maximum eye irritation,
at least where auto exhausts are irradiated. ’ 0 More details on the eye-panel selection and
testing are given in Appendix A.
Arithmetic and geometric mean response times and response indices are presented in the
supplementary Data Report. Individual responses of each panelist are also provided there for any
additional statistical analyses EPA may want to conduct.
In companng eye-irritation intensities between pairs of experiments in which aerosol
concentration was the only widely varying parameter (Table 1, Runs: AA-006 and AA-007,
AA-008 and AA-0l0, AA-016 and AA-017, and AA-028 and AA-029), it appears that the
amount of aerosol surface and/or volume has very little or no influence on eye irritation. To
more fully explore any possible relationship between these variables, linear regression equations
with two vanables were obtained (by method of least squares), and correlation analyses were
performed to describe the strength of the regression equations. Regression equations and
correlations were obtained with eye irritation (100/mean response time, see) as the dependent
variable and PAN (ppm), formaldehyde (ppm), and total aldehyde (ppm) as independent
variables in addition to aerosol surface (pm 2 /cm 3 ) and aerosol volume (pm 3 /cm 3 ) concentra-
tions. Values of the independent variables were taken near the times of the eye-irritation
measurements The results are shown in Table 2. As indicated by the correlation coeflicients iii
Table 2, there is virtually no correlation between eye irntation and aerosol or PAN concentra-
tions, and only moderate correlation between eye irritation and aldehyde concentrations In
conclusion, neither the concentration of aerosols nor the concentration of cci hun gaseous
products, taken individually, correlated highly with eye irritation associated with photochemical
smog arising from irradiation of typical hydrocarbon-NO and hydrocarbon-NO -SO 2 mixtures in
air.
Mechanisms of Aerosol Formation in Smog
Use of the MAAS has revealed several important features regarding aerosol formation in
smog which have been previously obscured because of inadequate aerosol-analyzing methods
Most important, perhaps, these measurements lead to a clearer understanding of aerosol growth
processes, which in turn enables US to speculate on the chemical mechanisms prevailing in aerosol
formation. Before discussing aerosol chemistry, it is instructive to describe the physical pattern
of aerosol formation as it occurs in a smog chamber operating as a batch reactor.
-------
11
TABLE 2. CORRELATIONS BETWEEN EYE IRRITATION AND VARIOUS INDEPENDENT VARIABLES
Independent
Variable
Linear Regression Parameter
Correlation
Coefficient
Number of
Observations
Slope
Intercept
Aerosol surface
concentration
26
—o oi
0.96
—0.08
Aerosol volume
concentration
26
—0.01
0.96
—0.12
Formaldehyde
concentration
24
0.55
0.75
0.40
Total aldehyde
concentration
24
0.21
0.62
0.58
PAN concentration
17
0.51
0.95
0.14
Physical Processes
Profiles of some of the parameters associated with aerosol formation during the irradiation
of filtered and diluted auto exhaust are presented in Figure 5. As indicated here, thcre is a very
rapid increase in the aerosol number immediately after irradiation is initiated. The very brief
induction period is the time required for the vapor of the incipient aerosols to reach cntical
concentrations at which nucleation begins. Once formed, the nuclei begin to coagulate. The
maximum in the total number of nuclei occurs when the rate of coagulation equals the rate of
nucleation. Beyond this point, the rate . of nucleation diminishes as more surface becomes
available to support condensation — thermodynamically a more favorable process. Thus, shortly
after the maximum in the total-number curve, condensation is believed to be the prevailing
mechanism by which aerosol volume increases. In this experiment, total aerosol production
(total-volume curve) increases linearly for about 1 hour and then diminishes. The decrease in the
volumetric conversion rate is generally attributed to (1) decreases in concentrations of precursor
constituents, (2) waIl losses, and (3) chamber dilution.
Increases in the concentration of total surface area result from nucleation and from
condensation of new vapor on existing aerosols. At the same time, however, coagulation reduces
the total aerosol surface. For the experiment depicted in Figure 5, after about 1 hour, the rate at
which surface is destroyed by coagulation is greater than the rate at which new surface is created
by condensation.
Integrated light scattering (bscat) is seen to increase throughout most of the irradiation,
leveling off only during the last hour. The fact that the slope of the total-volume curve after
2 hours of irradiation is less negative than the slope of the total-surface curve indicates that the
increase in light scattering after 2 hours is due primarily to coagulation of small aerosols
produced earlier in the reaction.
As mentioned previously, similar plots of total aerosol number, total volume, total-surface-
area concentrations, and integrated light scattenng versus irradiation time are available in the
supplementary Data Report for each experiment.
-------
12
U,
C
6O
FIGURE 5. PROFILE OF AEROSOL FORMATION DURING IRRADIATION OF FILTERED
AND DILUTED AUTO EXHAUST (16 PPM C HYDROCARBONS)
Chemical Processes
Many investigators have been under the impression that the formation of aerosols in smog us
directly related to the formation of ozone as evidenced by the correspondence in the appearance
of light scattering and ozone. Our findings indicate that such dependency is observed only for
o1efin-NO and olefin-NO -SO 2 smog systems, the latter resulting primarily in formation of
sulfate aerosol.
In aromatic hydrocarbon-NOr systems, this dependency is not observed and an alternative
dependence involving OH is suggested.
Oxidation of Hydrocarbons. When diluted auto exhaust is irradiated m the smog chamber.
most of the aerosol volume is produced during the period of NO to NO 2 conversion, as
illustrated in Figure 5. In an earlier study of the propensity of specific exhaust hydrocarbons to
generate photochemical aerosols, aromatic compounds stood out as being far more reactive than
most olefin or paraffin cornpounds. ’ 1) Thus the pattern of aerosol formation observed during
auto exhaust irradiation would be expected to appear during irradiation of aromatic
hydrocarbon-NOr-air mixtures.
100
80
40
2 3
Irrodiation Time, hr
-------
13
Figure 6 shows the formation of aerosol during irradiation of toluene and NO in air (Run
AA-008). In this case it is quite clear that aerosol production occurs almost exclusively during
the NO to NO 2 conversion period. The abrupt decline in aerosol production at 240 minutes
cannot be attributed to the consumption of toluene but may be associated with the consumption
of NO. When one examines kinetic data for reactions of aromatic hydrocarbons with the free
radicals in smog 11 2), it is clear that OH is the most important attacking species. Furthermore.
recent kinetic simulations of smog chemistry 1 ’ 3) reveal that, of the OH-forming reactions listed
below, the overall rate for Reaction (1) is much greater than that for all the others, unless, of
course, the NO concentration is very small.
700
600
E
I
4O0
a
C
C,
C.,
C
0
0
2O0 o
100
0
360
FIGURE 6. AEROSOL.FORMATION PROFILE DURING IRRADIATION OF
TOLUENE.NO AIR MIXTURE
Thus Reaction (1) serves to continually pump OH into the system to react with aromatic
hydrocarbons while also oxidizing NO to NO 2 . When the concentration of NO becomes very
small, other intermediates (e.g., 0 atoms, R0 2 RO, and HO 2 radicals) become competitive in
reacting with aromatics, and presumably the nature and/or rate of these reactions do not lead to
substantial aerosol yields.
HO 2 + NO ÷ HO + NO 2 (1)
HONO + hv + HO + NO (2)
H 2 0 2 + h 2H0 (3)
O(’ D) + H 2 0 + 2HO (4)
1.4
a
>
>
0.
0.
0.8
0
•1-
0
J06
02
120 I
Irradiation Time,min
RCHO 2 H RCHO + HO (5)
-------
14
Although most paraffins and low molecular-weight olefins yield much less organic aerosol
than aroniatics, olefins possessing 4 carbon atoms produce substantial quantities when irradi-
ated with NOR. For example, 1-heptene yLelds about 50 to 70 percent of the amount of aerosol
generated by equal concentrations of toluene.
The reaction between oleilus and ozone appears to be the most important in forming these
aerosols. As shown in Figure 7, substantial aerosol formation does not occur during NO oxida-
tion (where the O Il concentration is highest), but rather after NO is consumed and 03 is
generated. Thus, unlike the situation with aromatic hydrocarbons, aerosol formation froni the
higher-inoIecular- seight olefins seems to involve reactions with 03. Results of the dark reaction
of 03 with 1-heptene (Run AA-02 I) support the belief about the importance of 03 in the
irradiation experiment.
2 Oxidation. Numerous studies have been conducted to ascertain the interaction of SO 2
in photochemical smog, with particular concern for the formation of sulfunc acid aerosols.
Although it has been well established that sulfuric acid aerosols are indeed products of SO 2
oxidation, considerable uncertainty exists as to the rate of SO 2 oxidation and the prevailing
mechanisms involved. In most studies where SO 2 has been irradiated in air containing hydro-
carbons and nitrogen oxides, no quantitative results have been reported for SO 2 -oxidation
products. Thus it has not been clear as to the extent to which the observed losses of SO 2 in
these systems actually culminated in sulfune acid aerosols.
Recently Cox and Pcnket t14 conducted a series of dark reactions in which 503 was
successfully monitored by means of radiochemical techniques. Clark and Whitby ’ 5 determined,
indirectly, the rate of SO 3 formation from 502 oxidation by measuring aerosol products with
the same instrumentation employed in this study. in their experiments, SO 2 was irradiated in the
absence of other pollutants. Similarly in this study, irradiation experiments were carried out with
SO 2 added to air otherwise unintentionally contaminated (Runs AA-016 and AA-017), but also
to air to which hydrocarbons and N0 were intentionally added (Runs AA-015, AA-028, anti
AA-029). In the “50 2 -clean air” runs, humidity was the independent vanable.
Figure 8 shows the effect of humidity on the aerosol volumetric formation rate (dV/dt)
when 502 was irradiated alone in air- At 58 percent RH, maximum dV/dt was about three times
greater than that at 14 percent RH. It is of considerable interest to determine whether the
aerosol-formation rates observed in these experiments can be substantiated on the basis of SO 2
oxidation and aqueous sulfuric acid vapor-pressure data. ’ 6 If one assumes that the loss of SO 2
is due entirely to SO 3 formation, and further that equilibrium exists between the condensed and
vapor phases of aqueous sulfuric acid, SO 3 (H 2 OL, then the observed decay in SO 2 should be
equal to the observed production of S0 3 (H 2 0) aerosol. The appropriate calculation for
predicting the change in aerosol volume due to SO 2 oxidation is shown below for Run AA-016.
A [ S0 2 1kM
= xp (6)
where x = mole fraction of 50 3 H 2 O in SO 3 (H 2 0) at the temperature and humidity
of the system
p = density of S0 3 (H 2 O) at the same conditions
r = ratio of molecular weight of SO 2 to SO 3 H 2 O.
-------
>
>
E
a.
a.
C
0
0
.4—
C
C)
C .)
C
0
0
U ,
0
C
4
1.2
‘— 10
E
a
a.08
04
C)
02
0
15
700
600
‘ I ,
E
500
£
2001
100
FIGURE 7. AEROSOL-FORMATION PROFILE DURING IRRADIATION OF
1.HEPTENE-NO .AIR MIXTURE
In
E
U
t
C
0
4-
a
4-
C
C)
U
C
0
0
a)
E
0
U,
0
a ,
‘C
Irradiation Time, mm
FIGURE 8. EFFECT OF HUMIDITY ON THE VOLUME OF H 2 5O 4 AEROSOL
FORMED DURING IRRADIATION OF S0 2 -AIR MIXTURES
120 180
Irradiation Time, m m
0 60 120 ISO 240 300 360
-------
16
To simplify the calculation, the change in aerosol volume was predicted over the irradiation
interval of 0 to 60 minutes where rates were approximately linear. For the conditions of Run
AA-0l6, the constants in Equation (6) are: x = 0.40, p = 1.3 g/cm 3 , and r = 065. Actual SO 2
loss dunng the first hour (corrected for dilution) was 0.019 ppm. Using the units conversion
factor, I ppm SO 2 = 2610 pg/rn 3 (25 C, I atm), the calculated value for Vso 3 IH 2 o) at
60 minutes is 147 prn 3 /cm 3 . The observed value was about 160 pm 3 /crn 3 . Performing the same
type of calculation on results of Run AA-017 (where SO 2 = 0.015 ppm, x = 0.61. p =
1.5 g/cm 3 , r = 0.65) leads to a prediction for VSoa(H 2 o) of 66 pm 3 /cm 3 . The observed value
was 60 prn 3 /crn 3 . The good agreement in these instances supports the hypothesis that the
observed decay in SO 2 was indeed due to gas-phase oxidation of SO 2 to SO 3 rather than to
either absorption or heterogeneous reactions of SO 2 on chamber surfaces. What is difficult to
explain is the observed rate of SO 2 oxidation. The rates in Runs AA-0 16 and AA-0 1 7 ranged
from about 7 to 10 percent/hr. The maximum theoretical rate for conditions of this study is less
than 2 percent/hr. 1 6) One is therefore left with the conclusion that contamination is responsible
for the enhanced rate, but at contaminate concentrations <2 ppm C “nonreactive” hydrocarbon
and <0.05 ppm NOR. Clark and Whitby 5 also suspected that contamination may have
increased SO 2 oxidation rates observed in their work, although the oxidation rates they
calculated were within the theoretical limit, even when contamination was suspected.
Figure 9 shows the results of SO 2 oxidation and subsequent SO 3 (H 2 O) aerosol formation
when 0.06 ppm SO 2 was irradiated in air to which 1-heptene and NO were intentionally added
(The aerosol-volume profile is also indicated in Figure 9 for 1-heptene/NO irradiation in the
absence of SO 2 .) Unlike the results of irradiating SO 2 in clean air, there is a substantial delay
before reaching the maximum rate of aerosol formation, which is also coincident with the
maximum formation rate for ozone.
>
C
0
0.2
box
E
C
0
4-
0
4-
C
a)
U
C
0
0
4o
20
FIGURE 9. AEROSOL FORMATION PROFILE DURING IRRADIATION OF
1.HEPTENE.NO .AIR MIXTURES IN PRESENCE AND
ABSENCE OF SO 2
120
80
60
Irradiation Time, mm
-------
17
Several rnvcstigators, including Cox and Penkett 14 , argue that the most likely mechanism
for SO 2 oxidation under these conditions involves oxidation via zwitterions produced from
ozonolysis of l-heptcne. A reasonable extension of the theory is the presumption that the SO 2
oxidation rate, ou SO 3 (l-1 2 O), formation rate, would be first order in each of three reactants
(S0 2 ,0 3 , and 1-heptene) as indicated by Equation (7).
d SO 3 (l-l 2 O) /dt = k (SO 2 ) (0 3 )(l-heptene) . (7)
During the period from 90 to 150 minutes where d SO 3 (H 2 O) /dt was nearly constant, the
product of the concentration terms would be expected to be constant. Inserting the appropriate
values of SO 2 , 03, and 1-heptene at irradiation times of 90, 120, and 150 minutes yields
products of 0 38, 0 46, and 0.40 ppm 3 , respectively, and the first-order assumption appears
valid. However, the abrupt decrease in dV/dt near 200 minutes severely contradicts predictions
based on this first-order scheme. At 200 minutes, the [ 03] [ S0 2 ] [ olefm] product was
0.23 ppm 3 , which would indicate an aerosol-formation rate of about half the maxinium (4
.zm 3 cm- 3 mm-’), and not zero. Reasons for the abrupt change in slope of the aerosol-volume
curve at 1 80 to 200 minutes are not apparent. It was not due to aerosol losses because at this
time the mean aerosol diameter was about 0.1 ,im — a size that is not especially susceptible to
wall loss due to diffusion or settling.
The maximum rate of SO 2 oxidation in Run AA-028, whether based on SO 2 decay or
aerosol production, was about 7 percent/hr, or less than that observed in the absence of the
olefin and NO pollutants. It appears then that there is no evidence (here) to support or deny
the zwitterion theory for SO 2 oxidation in irradiated atmospheres. In the zwitterion theory,
there is nothing to explain the SO 2 oxidation rate where very little olefins and 03 could be
detected, or the abrupt decline in SO 2 oxidation where substantial amounts of olefin, SO 2 , and
03 were still present. In authentic smog atmospheres, homogeneous SO 2 oxidation may be very
complicated. Energy transfer from important intermediates might play a significant inhibitory
role. Zwitterions may well be important intermediates for SO 2 oxidation, but it is also likely
that other constituents of smog are competitive with SO 2 for reactions with the diradical.
Influence of Primary Auto-Exhaust Aerosols on
Development of Secondary Aerosols
To facilitate experimental control, most smog-chamber studies have been conducted with
particle-free air — a condition which is obviously unrealistic compared with urban atmospheres.
Results from APi project EF-2 at BattelIe 1 1) indicated that primary automobile particulate had
a positive effect on the amount of light scattering that developed during irradiation of
automobile-exhaust vapors. The changes in light scattering could not be clearly interpreted,
however, because techniques were not available for measuring the concentration of suboptical-
size particles. In the present program, the effect of preexisting particles on photochemical-aerosol
development was reinvestigated, using the MASS in conjunction with auto-exhaust experiments.
The concentration of primary aerosol was controlled by filtering the exhaust while adding it to
the smog chamber. In some experiments, exhaust was added without filtering. Two series of
experiments were conducted in this manner — one at higher primary aerosol concentrations than
the other. The different sources of auto exhausts, automobile operations, and methods of
transferring exhaust to the smog chamber are described in Appendix A.
-------
18
Moderate Versus Zero Primary-Aerosol Levels. In the first set of expenments, conducted
with an APi project EF-8 automobile, carburetor enrichment was used to provide a suitable
exhaust HC/NOX ratio for smog-chamber experiments. This enrichment also provided larger than
normal amounts of primary aerosol, such that the total-surface-area concentration of primary
aerosol present in the chamber at the outset of irradiation ( l300 Mm 2 cm- 3 ) approached that
existing in ambient air in the mornings as photochernical reactions begin.
Results of irradiating auto exhaust in the presence (unfiltered) and absence (filtered) of
primary aerosol are shown in Figure lO.* With the exception of the differences in primary
aerosol concentration, the experiments were essentially identical, and the ensuing gas-phase
reactions which were monitored were also very similar.
100
80
60
40
20
U,
C
‘a-
0
C
w
U
I.-
G)
FIGURE 10. EFFECT OF PRIMARY AUTO-EXHAUST PARTICLES ON SECONDARY AEROSOL
GROWTH AND LIGHT SCATTERING (16 PPM C HYDROCARBONS)
The two aerosol-volume-concentration curves in Figure 10 indicate that initially the volu-
metric conversion rates are nearly the same and that, at the end of the irradiation, the difference
in total volume is nearly equal to the initial difference, i.e., the volume of primary exhaust
aerosol. Therefore, the volume of aerosol formed photochemically is not substantially influenced
by the presence of primary aerosol and the large difference occurring in integrated light
scattering must be associated with other properties of the aerosols. As indicated in Figure 10, the
These data and discussion are included in the Proceedings of the Third International Clean Air Congress (Reference I) .ind in
Battelie-Columbus’ Interim Report on “Motor Fuel Composition and Photochemicai Smog” to the American Petroicum
Institute (Reference 10).
2 3
Irradiation Time, hr
-------
19
initial formation rate of aerosol surface was significantly greater for the filtered exhaust. This
observation is consistent with the difference in nucleation rates (not shown in Figure 10); more
nuclei were formed and therefore more surface was initially formed in the case of filtered
exhaust. As the irradiations progressed, the total surface area concentrations became similar after
nucleation subsided.
Light scattering, however, does not depend on the total amount of surface area associated
with particles, but rather on the amount of surface area associated with particles >0.2 urn in
diameter (not a strict cutoff at 0.2 Mm) and increases as the diameter of the particles approach
0.5 urn. Thus the differences observed in light scattering can be explained best by examining the
distribution of surface over the aerosol size range.
The distributions at irradiation periods of 0.5 and 4 hours are indicated in Figure II for the
two auto-exhaust experiments discussed above. (The surface distributions in Figure 11 are
dimensioned such that the area under a curve in a given size range is directly proportional to the
surface area in that size range.) Upon comparing the surface distribution of the aerosol in the
filtered- and unfiltered-exhaust experiments at 0.5 hour, one notes that the total area under the
curve in the filtered case is greater than that in the case where particles were present initially.
However, the area under the surface curve in the important light-scattering range (>0.2 Mm) is
much greater for the unfiltered exhaust. Comparison of the results at 4 hours reveals that the
amount of surface associated with aerosols in the light-scattering range is again greater for the
unfiltered exhaust and that the distribution approaches the size range of optimum scattering
efficiency.
In
E
U
( J
E
a
a’
0
0
U,
0
FIGURE 11. EFFECT OF PRIMARY AUTO-EXHAUST PARTICLES ON THE SURFACE
DISTRIBUTION OF SECONDARY AEROSOLS (IRRADIATED AUTO
EXHAUST. 16 PPM C HYDROCARBONS)
Particle Diameter (Dr), m
-------
20
Thus the differences in light scattering observed upon irradiation of filtered and unfiltered
auto exhaust can be explained primarily by differences in the resulting aerosol surface distribu-
tions rather than by differences in total volume of photochemical aerosols generated.
Low Versus Zero Primary Aerosol Levels. Another series of exhaust experiments w.ts
conducted using two similar automobiles, one operating on leaded fuel and the other on
nonleaded fuel. The history of these two automobiles, provided by the EPA/CRC Project
CAPE-19 committee, is well documented. t18 For these experiments, suitably high HCINOX
ratios were obtained by adding a synthetic mixture of auto-exhaust hydrocarbons to actual
exhausts emitting from the cars operating in their normal configuration. The ratio of synthctic/
actual exhaust was 5/3. Because the cars were operated normally and the actual exhaust was
diluted with synthetic exhaust, the amounts of primary aerosol present before irradiation were
much less than in the preceding experiments using the API automobile. With unleaded fuel, the
initial aerosol surface-area concentration was only ‘50 pm 2 cm - 3 , with leaded fuel, it was
‘-‘SOO pm cm- 3 . Because of the small concentration of primary aerosol, no substantial differ-
ences in the size distributions and volume concentrations were observed in comparing the
filtered- and unfiltered-exhaust experiments
As indicated in Table 1, there was no apparent difference in the photochernical reactivity of
exhausts derived from leaded or nonleaded fuel. Respective rates of NO oxidation, maxiinuin 03
and aldehyde concentrations, and the degree of eye irritation were also similar for both types of
exhaust.
In conclusion it appears that preexisting (primary) aerosols have little ellect on the volume
of aerosol matenal produced by photochemical processes. However, at aerosol surface concentra-
tions similar to those in the atmosphere, suboptical-size aerosols effectively increase the sue of
the aerosols formed secondanly, which is likely to increase the total amount of light scattering
and thus account for observations reported in the EF-2 program.t 1 1)
REFERENCES
(1) Miller, D. F, and Levy, A., “Aerosol Formation in Photochemical Smog The Effect of
Humidity and Small Particles”, presented at the Third International Clean Air Congress,
Dusseldorf, Germany (October, 1973).
(2) Whitby, K. T., Clark, W. E., Marple, V. A., Sverdrup, G. M., Willeke, K., LuLl, B.Y.H., and
Pui, D.Y.H., “Evolution of the Freeway Aerosol”, presented at the ACS Annual Meeting,
Chicago, Illinois (August, 1973).
(3) Whitby, K. T., Wilson, W. E., Pui, D.Y.H., Miller, D. F., Levy, A., Kittelson, D. B., and
Cantrell, B K., “Effect of Stirring on Aerosol Production in a Smog Chamber”, presented
at the ACS Annual Meeting, Chicago, Illinois (August, 1973).
(4) Miller, D. F, Wilson, W. E., Pui, D.Y.H., Whitby, K. T., and Levy, A., “The Size
Distribution and Concentration of Combustion and Photochemical Aerosols Attributable to
Automobiles”, paper No. 74-1 5 I, presented at the 67th Annual Meeting of the Air
Pollution Control Association, Denver, Colorado (June, 1974).
-------
21 and 22
(5) Scofield, F., Levy, A., and Miller, S. E., National Paint, Varnish and Lacquer Association,
Inc., Circular No. 797 (January 3 1969).
(6) Tuesday, C. S., “Chemical Reactions in the Upper and Lower Atmosphere”, lnterscience,
New York, New York (1961).
(7) Pitts, J. N., Jr., Vernon J. M., and Wan, J.K.S., Intern. J Air and Water Poll. 9, 595-600
(1965).
(8) Gordon, R. J., “Pilot Study of Ultraviolet Radiation in Los Angeles”. J. S. Nader (Ed.),
National Air Pollution Control Administration, Durham, North Carolina (1967).
(9) Whitby, K. T., Liu, B.Y.H., Husar, R. B., and Barsic, N., Jr., “The Minnesota Aerosol-
Analyzing System Used in the Los Angeles Smog Project”, J. Colloid Interface Sci., 39, 136
(April, 1972).
(10) Levy, A., Miller, D. F., Hopper, D. R., Spicer, C. W., and Trayser, D. A., “Motor Fuel
Composition and Photochemical Smog”, Battelle-Columbus Interim Report to the American
Petroleum Institute on Project EF-8 Report No. CEA-4 (December, 1973)
(11) Wilson, W. E., Jr., Miller, D. F., Levy, A., and Stone, R. K., J Air Poll Control Asso,
23(11), 949-56 (1973).
(12) Morris, E., and Niki, H., J. Phys. Chem., 75, 3640 (1971).
(13) Demcrjian, K. L, Kerr, J. A., and Calvert, J. G., Adv. in Environ Sci and Tech, 4 (1973)
(14) Cox, R. A., and Penkett, S. A., Trans. Faraday Soc., 1, 68, 1735 (1972).
(15) Qark, W. E., and Whitby, K. T., “Measurement of Aerosols Produced by the Photochemical
Oxidation of SO 2 in Air”, Particle Laboratory Publication No. 181, University of Minnesota
(July, 1972).
(16) Perry, J. H., Chemical Engineer’s Handbook, 4th Edition, McGraw-Hill, New York (1963),
pp 3-79.
(17) Sidebottom, H. W., Badcock, C. C., Jackson, G. E., Calvert, J. G., Reinhardt, C. W., and
Damon, E. K., Envi. Sd. & Tech., 6(l), 72 (January, 1972).
(18) Melton, C. W., Mitchell, R. 1., Trayser, D. A., and Foster, J. F., “Chemical and Physical
Characterization of Automobile Exhaust Particulate Matter in the Atmosphere”,
Battelle-Columbus Final Summary Report to the Coordinating Research Council (CAPE
19-70) and the Environmental Protection Agency (Contract No. 68-02-0205) (June, 1973).
-------
A-i
APPENDIX A
EXPERIMENTAL METHODS
Automobile Exhaust Generation and Sampling
Chassis Dynamometer Operations
The chassis dynainomcter used in this program is a Clayton Model C’T-200 with a variable-
inertia flywheel assembly capable of simulatmg vehicle weights from 1500 to 5500 pounds
Automobiles were operated by an automated speed controller actuating servomotors to the brake
and accelerator. The speed controller functions as a closed-loop system that varies vehicle speed
to match a speed-versus-time program prerecorded on magnetic tape.
A slightly modified version of the 1972 Federal driving schedule was used for the experi-
ments. This driving cycle, often referred to as the LA-4 cycle, consists of 22.8 minutes of
acceleration, cruising, deceleration, and idle modes covering 7.5 miles. The Federal Test Pro-
cedure requires the test car to be started cold from an overnight soak (at laboratory tempera-
ture) and almost immediately started into the cycle. To minimize the danger of engine stall
during a run, this procedure was modified by including a preliminary “prep” cycle before the
actual exhaust-sampling cycle, and to cool the engine and exhaust system for 45 minutes using
an external coolant heat exchanger, an exhaust-system cooling blower, the radiator fan, and a
cooling air stream on the choke box, carburetor, and intake manifold. During this 45-minute
rapid cooldown, all measured temperatures, with the exception of the oil-sump temperature,
were decreased to 80 F or below
Vehicles and Fuels
Exhausts irradiated in Experiments AA-024 and AA-025 were generated from a 1971
Chevrolet used concurrently on an API program (Project EF-8) at Battelle-Columbus. The API
agreed to supply the automobile and fuel for these expenments in exchange for a report on the
results. The Chevrolet, a BelAir 4-door sedan equipped with a 350-CID V-S engine and two-barrel
carburetor, had been operated only on unleaded fuel. The original HC/NO ratio of the car’s
exhaust was too low for practical smog-chamber experimentation. Therefore, a “California
camshaft” was installed and the main carburetor jets (0.058-inch bore) were replaced with jets of
larger bore (0.067 inch). As documented in the EF-8 program t ‘° , these modifications had little
or no effect on the hydrocarbon composition of the exhaust, but increased the total hydro-
carbon, CO . and aerosol emissions. Fuel used in both experiments was API No. 13A (nonleaded)
consisting of 5) percent paraffins, 13 percent olefins, and 32 percent aromatics by weight.
For experiments AA-032 through A.A-038, two similar 1970 Fords were used for generating
exhaust from leaded and nonleaded fuel. The history of these cars has been carefully monitored
for comparing leaded- versus nonleaded-fuel effects on primary particulate emissions. 1 8) Both
cars are Fairlane two-door hardtops equipped with 35 1 CID V-8 engines, and two-barrel
carburetors. Operation of these cars on the chassis dynamometer was the same as that described
-------
A-2
above for the API vehicle. The CRC-CAPE-19 fuels, with and without TEL additive (2.45 g
Pb/gal), consisted of 63 percent paraftins, 6 percent olefins, and 29 percent aromatics by weight.
No modifications were made to these cars to increase the HC/NO ratio. lnstead, synthetic-
exhaust hydrocarbons were added to the actual exhaust to increase the ratio for the smog-
chamber experiments.
Sample Transfer
The goal in auto-exhaust sampling is to produce and transport to the smog chamber an
emission sample reasonably representative of vehicle emissions on roadways The exhaust-
sampling facility used on this program is shown schematically in Figure A-I. An exhaust-diverter
valve (between the tailpipe of the car and the dilution tunnel) controls the period when the
exhaust gases are passed through the tunnel. The dilution tunnel was designed to rapidly quench
and thoroughly mix the exhaust gases with clean filtered air, and to establish a flat and
reproducible velocity profile at the sampling position. The tunnel is constructed of stainless steel
(11-1/2 niches in diameter and 20 feet long), and includes a charcoal filter, an absolute filter,
and a mixing orifice at one end and a sampling tube and CVS (constant-volume-sampling) system
at the other end. The CVS system induces the dilution-air and exhaust-gas flow through the
tunnel. The unit is an Olson Laboratones Model No. 45A-R3 with a four-speed pump motor.
During all exhaust-sampling runs the total flow rate was 340 scfm. The volume flow rate of
exhaust into the tunnel varied from 15 to ISO cfm. The average dilution-air/exhaust-gas ratio
over the entire driving cycle is about 10/I. In an effort to increase this ratio, thereby niinimi7ing
the coagulation rate of primary aerosols, an exhaust splitter was installed which diverted only
one-fourth the exhaust into the tunnel. This dilution procedure (yielding an average ratio of
40/I) was used on Runs AA-033, AA-035, and AA-036.
As illustrated in Figure A-l, a portion of the diluted exhaust is delivered from the near end
of the tunnel to a Teflon holdmg bag (550-cu ft capacity) contained in a rigid box. During
collection, near-isokinetic flow from the tunnel to the bag is achieved by maintaining a constant
subatmospheric pressure on the bag. At the conclusion of the driving cycle, the contents of the
bag are passed to the smog chamber by applying super-atmospheric pressure on the bag.
FIGURE A-i. SCHEMATIC OF LABORATORY FACILITY FOR GENERATING
AND SAMPLING AUTO EXHAUST
Project car
Exhaust
chassis
dynamomel er
line
Pressure or
vacuum tine To smog
chamber
-------
A-3
For experiments where the total exhaust flowed into the dilution tunnel (Runs AA -024,
AA-025, AA-032, AA-034, AA-037, and AA-038), rapid coagulation of primary aerosol was
forestalled by filling the bag half full of filtered air before introducing the exhaust emission.
Despite only moderate increase in overall dilution (20/1 rather than 10/1), this technique is
fairly successful because the emission associated with the first few minutes of the driving cycle
and containing the greatest amount of aerosol is initially diluted manyfold.
Obviously, actual roadway operation of an automobile provides dilution of exhaust more
rapidly than can be provided practically in the laboratory, where constant volume sampling and
other experimental constraints are imposed. At the overall dilution ratio of 20/1, the cumulative
coagulation of primary aerosols (nonleaded fuel combustion) occurring during the 23-minute
driving cycle resulted in a unimodal distribution of aerosol surface, with the mode occurring at a
particle diameter of 0.1 1Am.
Diluting 40/1 in the dilution tunnel resulted in a bimodal distribution of aerosol surface,
the major mode occurring at a particle diameter of 0.03 pm and a minor mode at about 0.3-pm
diameter.
Dr. Whitby and his colleagues have recently demonstrated that the major mode in the
surface distribution of primary auto-exhaust aerosols having undergone dilution along otherwise
unpolluted highways is near 0.02 pm. 2 Size-distribution measurements (MAAS) taken directly
in Battelle’s dilution tunnel confirm that the initial mode in the size distribution is near 0.02
pm, and that the growth in particle size upwards from 0.02 pm is due to the coagulation of
aerosols dunng the collection period. The lifetime of these very small aerosols in urban
atmospheres and in heavily trafficked areas is uncertain and will depend on many factors,
including particle concentration, turbulence, and the rate of photochemical activity.
Eye-Irritation Measurements
Measures of eye irritation were made by a 24-member eye-panel team chosen from
Battelle-Columbus staff. Panelists selected were distributed among three 7-member teams (plus an
alternate) according to their responses to three chemical systems:
(1) Clean air (with lights on during eye test)
(2) 2-butene (4 ppm), NO (1.1 ppm), NO 2 (0.9 ppm) - irradiated
(3) Toluene (4 ppm), NO (1 ppm), NO 2 (1 ppm) — irradiated.
The last two systems were selected to expose the panelists to PAN and presumable PbzN,
respectively, as well as to significant levels of formaldehyde. The formaldehyde levels in these
two reactions were 0.7 ppm in the trans-2-butene system and 0.1 ppm in the toluene system.
Response times of the panelists were as follows:
Clean Air Trans-2-Butene Toluene
Mean Time, equivalent sec 342 66 82
Index 0.28 7.8 7.5
-------
A-4
The mean time in equivalent seconds is defined as the antilog of the geometric-mean response
time. The index is defined as l0 [ (360-t)J, where t is the response time. Maximum exposure time
was 360 seconds.
For all experiments with auto exhaust, two teams (14 panelists) were exposed during the
fifth hour of irradiation. For most experiments with specific hydrocarbons and synthetic
exhaust, panelists were exposed about 2 hours after the time of maximum NO 2 concentration.
One team of panelists was seated for all experiments. The other two teams alternated froii
experiment to experiment.
Detailed data of all eye-panel measurements made in this program are presented in the
supplementary Data Report. An example of the eye-data tabulation for one experiment (AA-
032) is shown below.
RU ’ AA-1132
5/ ?/73 (3: 15) 1U L E—l4l6
LYE 1scR [ IA! 1¼3 1’J hESP 4SES AND INDICES
NAME
N I-’ T L3C (T)
INL)EX SEvEt I F’i
A IN0r (
C 1EA’ )
T.DEV.Ci iLtC i),S
ST. LkiJ. 1’, J : ‘L ) , 5/ 53. [ Si. (1 ‘I)
EC’JI’v’ L NT C Nfl J
MEA’.i
‘i iEN 3T [ S
I r—.I 1 •
ii: ..i i •. ,
1 )E’J JTL.S
L) ( I)
I’)! ’. 4 =
. VE..I ? I
i 4’:; L! SI
i : i
•VSi )N 5
i F ES
i •-
‘ 0E .< JZ “i
I’)SI iIJJ .‘\T 1HE CHt’ i ’ fl’
TIME. L’J S!CJNDS
1) Li(: •)1 I ES J 4• j [ ’4•.’
- T’ 3 0
i .J—3 .4lIN i i:•: it f 5 I .\iI j t 1 1•
\‘ .L T
L)LCK CLA!*<
‘ALi CH.\PIN
J )E EJII
E3t)rl C0’JLØ.’J
T?AY DJ01Th
CriAI LE.5 0U [ L ’
‘ Ai GAi )
!3 )i3 LIT1L. :
> 4; (3A, Ei flELL
kAY ii\ ,CEi f
H I (:Rr’a, 1 J
‘IILL < I (Lt Y
, A:JLI ‘JI•.. fl\tJ’
i L i . 1 I I i C.<
(;)>1,’JT ’\ I I3.’
11)
15
13
16
2
4
1
23
2!
ii
Ij 52.
1 143.
2 207.
5 190.
C) 360.
3 35.
—
I r .jj.
I; H)5.
7 124.
Ii 236
2 157.
5 141.
3 4 ) .
1 124.
151.
—
) f
23.
56
6. u3
‘i.25
zl. 7 )
(1. 00
9. 0..3
. :31
/. u:
6. :D6
3. j1i
5. 54
1. US
•
..
. 79
2. 110
‘‘. 64
I • 71 6
2. 1 55
2.316
2 • 2 79
2.556
I • 544
2.312
2.021
2.093
2.3 7 4
2. 196
2. 1 119
1. 623
2. (19:3
2. 10
0.29
0. 0
1 2 6.
2
2
0
J
I • ‘1
-------
B-i
APPENDIX B
SUMMARY OF AEROSOL COLLECTIONS
TABLE B-i. SUMMARY OF AEROSOL COLLECTIONS BY FILTRATION
Sampling
Sample
Weight
Time,
Volume.
Gain,
Run Sample Description(a) Type of Filter mm
ft 3
mg - Dispositmon(b)
AA-005 Toluene-NO 4-in, glass — 388 10.1 BCL
AA-006 Toluene.NO 4-in, glass — 353 11.1 BCL
AA-007 ToIuene-N0 4.in. glass — 392 3 1 BCL
Gold disk 25 8.8 — Ditto
1 -in. glass 25 8.8 —
AA-008 1.heptene.NO 4-in, glass — 370 1 9 BCL
AA.009 1.heptene-NO 4-in, glass — 382 0.7 BCL
AA-010 1-heptene.NO 4-in, glass — 315 0.7 BCL
AA-011 Benzene-N0 4.in. glass — 338 1.1 BCL
AA-012 Toluene.mesitylene.NO 4-in, glass — 360 3.4 BCL
Gold disk 30 10.5 — Ditto
1-in glass 30 10.5 —
AA-U15 Benzene-NO -SO 2 4-in, glass — 374 1 5 BCL
AA-016 SO 2 4-in, glass — 374 1.8 BCL
AA-019 Propylene-0 3 -S0 2 -dark 4-in, glass — 435 3.2 BCL
Nuclepore 42 6 — Ditto
AA-020 -Pmnene-NO Nuclepore 42 6 BCL
Gold disk 30 7 1 Ditto
1-in, glass 30 7.1
AA-021 1-heptene-0 3 -dark 4-in, glass — 430 08 BCL
Nuclepore 27 10 — Ditto
Gold disk 30 10.5 —
1-in glass 30 10.5 —
AA-022 1-heptene-0 3 -S0 2 -dark 4-in, glass — 500 4 0 BCL
Nuclepore 28 10 — Ditto
Gold disk 30 10.5 —
1-in, glass 30 10.5 —
AA-023 Toluene-NO -CN 4-in, glass — 346 BCL
Nuclepore 30 10 Ditto
AA-024 Auto exhaust — smog 4-in, glass — 374 1.7 BCL
Nuclepore 30 10 — Ditto
Silver 26 10 —
Auto exhaust — primary Nuclepore 22 22 —
Silver 22 22 —
AA-025 Auto exhaust — smog 4-in, glass — 385 1,4 BCL
Nuclepore 7 7 — Ditto
Silver 7 7 —
Auto exhaust — primary Nuclepore 30 10 —
Silver 22 11 —
-------
B-2
TABLE B-i. (Continued
Sampling
Sample
Weight
Time,
Volumes
Gain.
Run Sample Description(a) Type of Filter mm
ft 3
mg
AA-028 1.heptene.NO -SO 2 4-in, glass 46 313 0.9 BCL
Nuclepore 30 10 — EPA
1-in, glass 25 9 — BCL
Millipore 53 — 0.004 EPA
Silver 16 10 — EPA
AA-029 1-heptene-NO .SO 2 4-in glass 47 317 0.75 BCL
Nuclepore 66 21 — EPA
l.in. glass 46 16.5 — BCL
Mullipore 55 — —0.002 EPA
Silver 17 10 — EPA
AA-030 1-heptene-NO 4-in, glass 53 342 0.5 BCL
Nuclepore 63 21 — EPA
1-in, glass 30 10.8 — BCL
Millipore 61 — 0.007 EPA
Silver 17 10 — EPA
AA-031 1-heptene-NO -CN 4-in, glass 51 334 0.59 BCL
Nuclepore 56 20 — EPA
1-in, glass 28 10 — BCL
Millipore 60 — 0.004 EPA
Silver 20 11 — EPA
AA.032 Auto exhaust-CN 4.in. glass 52 338 0.56 BCL
Nuclepore 56 20 — EPA
1-in, glass 31 10 — BCL
Millipore 62 — 0.022 EPA
Silver 17 10 — EPA
AA-033 Auto exhaust 4.in, glass 51 327 0.54 BCL
Nuclepore 56 20 — EPA
1-in.glass 31 11 — BCL
Millipore 66 — 0015 EPA
Silver 16 10 — EPA
Before irradiation Ab-Nuclepore — 2 EPA
After irradiation Ab-Nuclepore — 5 EPA
AA.034 Auto exhaust — filtered 4-in, glass 52 355 0 47 BCL
Nuclepore 55 20 — EPA
1-in glass 26 9 — BCL
Millipore 65 — 0005 EPA
Silver 16 10 — EPA
AA-035 Auto exhaust (leaded) 4-in, glass 56 348 0 42 BCL
Nuclepore 57 25 — EPA
1-in, glass 34 12.2 — BCL
Millipore 65 — 0.005 EPA
Silver 16 10 — EPA
-------
B-3
TABLE B-i. (Continued)
Run
Sample Description(a)
Type of Filter
Sampling
Time,
mm
Sample
Volume,
ft 3
Weight
Gain,
mg
o spos 1 t 1 o n(b)
AA-036
Auto exhaust (leaded)
Before irradiation
4-un. glass
Nuclepore
1-in, glass
Millipore
Silver
Ab-Nuclepore
50
46
38
66
26
—
334
20
13.6
—
12
2
0.36
—
—
0.017
—
—
BCL
EPA
BCL
EPA
EPA
EPA
After irradiation
Ab-Nuclepore
—
5.5
—
EPA
AA-037
Auto exhaust (leaded) — filtered
4-in, glass
Nuclepore
1-in, glass
Millipore
Silver
61
74
32
62
20
369
25
11.5
—
10
0.43
—
—
0.022
—
BCL
EPA
BCL
EPA
EPA
AA-038
Auto exhaust (leaded) — filtered
4-in, glass
Nuclepore
1-in, glass
Millipore
Silver
49
71
32
64
18
349
24
11.5
—
10
0.45
—
—
0.008
—
BCL
EPA
BCL
EPA
EPA
(a) Experimental conditions Uniess otherwise indicated, au aerosol samples were withdrawn from the smog chamber after
irradiation
(b) EPA. fiiters mailed to Project Officer at EPA BCL, fuiters retained at Batteiies Columbus Laboratories
------- |