RESEARCH REPORT
HAZE FORMATION --
ITS NATURE AND ORIGIN
to
ENVIRONMENTAL PROTECTION AGENCY
CPA 70-Neg. 172
and
COORDINATING RESEARCH COUNCIL
CAPA 6-68
O Batteile
Columbus Laboratories
-------
BATTELLE'S COLUMBUS LABORATORIES comprises the origi-
nal research center of an international organization devoted to research
and development.
Battelle is frequently described as a "bridge" between science and
industry a role it has performed in more than 90 countries. It
conducts research encompassing virtually all facets of science and its
application. It also undertakes programs in fundamental research and
education.
Battelle-Columbus - with its staff of 2500 - serves industry and
government through contract research. It pursues:
research embracing the physical and life sciences, engi-
neering, and selected social sciences
design and development of materials, products, processes,
and systems
information analysis, socioeconomic and technical eco-
nomic studies, and management planning research.
505 KING AVENUE COLUMBUS, OHIO 43201
-------
HAZE FORMATION
ITS NATURE AND ORIGIN
to
ENVIRONMENTAL PROTECTION AGENCY
CPA 70-Neg. 172
and
COORDINATING RESEARCH COUNCIL
CAPA 6-68
by
Wm. E. Wilson, Jr., Warren E. Schwartz,
and G. W. Kinzer
January 28, 1972
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
TABLE OF CONTENTS
Page
INTRODUCTION 1
SUMMARY AND CONCLUSIONS 1
EXPERIMENTAL 5
SAMPLING SITES 5
Bronx Site. 5
Cooper Union Site 6
Blue Ridge Mountain Site. ...... 7
COLLECTION AND GENERATION OF AEROSOL 7
ORGANIC ANALYTICAL PROCEDURES . . 9
RESULTS AND DISCUSSION ........ 9
CYCLONE DESIGN AND PERFORMANCE EVALUATION . 9
LIGHT SCATTERING STUDIES. 15
ORGANIC ANALYTICAL STUDIES. .... 19
Infrared Spectroscopic Analysis of Total Particulate . 19
Analysis of Particulate Extracts: Sample
Sources and Analytical Objectives 25
Comparison of Air Particulate and Auto
Exhaust Particulate 28
Chemical Fractionation and Infrared
Spectroscopic Analysis. .. 29
Analysis by Gas Chromatography and Gas
Chromatography Interfaced with
Mass Spectrometry 31
Analysis of Smog Chamber Aerosol 33
BATTELUE COLUMBUS
-------
TABLE OF CONTENTS
(Continued)
Page
Aerosol Generated from Cyclohexene 39
Aerosol Generated from a-Pinene 45
Aerosol in Natural Haze 54
Aerosol Generated from 1-Heptene. . 54
Aerosol Generated from Toluene 62
Cross-Contamination in Generation
of Smog Chamber Aerosol 62
Summary: Organic Reactions in
the Smog Chamber. . . . 63
Analysis for Polynuclear Aromatic Hydrocarbons .... 64
INORGANIC ANALYTICAL STUDIES 69
Bronx Site 69
Smoky Mountains Site. . 69
ACKNOWLEDGMENT . . 71
ATTELLE COLUIVIBUS
-------
FINAL REPORT
on
HAZE FORMATION -- ITS NATURE AND ORIGIN
to
ENVIRONMENTAL PROTECTION AGENCY
CPA 70-Neg. 172
and
COORDINATING RESEARCH COUNCIL
CAPA 6-68
from
BATTELLE
Columbus Laboratories
January 28, 1972
INTRODUCTION
The ultimate goal of this program is the development of a field
technique capable of elucidating the contribution of automobile exhaust
to atmospheric haze. To attain this goal, it is necessary to establish
the relationship between atmospheric haze and automotive emissions and to
discover methods for distinguishing between the man-made and natural hazes.
Experimental studies involve measurements of the optical and chemical
properties of the atmospheric aerosol and determination of the gaseous
pollutants in the atmosphere in which the aerosol is found. As a result
of the first year's study, significant insight has been gained in under-
standing the nature and magnitude of the problems inherent in specifying
the contributing sources for haze.
SUMMARY AND CONCLUSIONS
The objective of the first year's program was to determine if
the organic composition of atmospheric aerosols could be used to identify
BATTELLE COLUMBUS
-------
their sources. Particulate associated with the following sources was
collected for analysis:
(1) urban and rural atmospheric haze
(2) primary auto exhaust
(3) aerosol generated from known precursers
in the Battelle-Columbus smog chamber.
Field sampling of haze particles was performed at two locations
in New York City and in the Blue Ridge Mountains. Samples were collected
at the Bronx State Hospital grounds during the full month of September, 1970.
From September 21, through October 2, 1971, samples were also collected in
mid-Manhattan on the rooftop (fifth floor) of the Cooper Union BuildfLng.
Battelle's mobile laboratory was then moved to the Blue Ridge Mountains,
where continuous sampling was conducted from October 12 to October 23, 1970.
The collections of air-borne particles were made with two hi-vol
samplers, one conventionally designed and one equipped with a cyclone,
(see Cyclone Design and Performance Evaluation) which "cut out" nonre-
spirable and nonlight-scattering particles (> 2 microns). Samples were
obtained on both glass and nylon fiber filters.
At all three sites, simultaneous measurement of visibility
reduction was made with an integrating nephelometer. Concurrent measure-
ments of the gas phase components were expected to have been made at the
Bronx site by Scott Laboratories, but scheduling precluded the collection
of most of those data. Monitoring in the Blue Ridge Mountains included
sunlight intensity, relative humidity, ozone concentration, condensation
nuclei count, and light scattering with a multiwavelength integrating
nephelometer.
Histographs made of the light scattering data from the sampling
site.s showed that the average visibility during the sampling periods was
40 miles for the Blue Ridge Mountain site, 12 miles for the Bronx site,
and 9 miles for the Manhattan site. The most severe visibility conditions
recorded at the respective sites were 14, 5, and 3 miles.
The detection of a terpene oxidation product in the Blue Ridge
Mountain aerosol and the relationships observed between light scattering,
sunlight, and ozone formation confirmed the notion that the blue haze is
a photochemical aerosol fueled in part by the terpene emissions from trees.
BATTELLE COLUMBUS
-------
The engine-exhaust aerosol was generated from a 1967 Chevrolet
operated on a chassis dynamometer. The exhaust from one 7-mode Federal
driving cycle was vented into a large (300 cu ft) Mylar bag containing
sufficient dry nitrogen to prevent condensation of water vapor. The
engine aerosol (primary) was collected by evacuating the bag through a
glass-fiber filter.
The aerosols generated from individual hydrocarbons were pre-
pared in a 610-cu-ft environmental chamber by irradiating 10 ppm of
selected hydrocarbon and 1 ppm of NO. in clean, humidified (50 percent
RH) air. Measurements of visibility reduction and oxidant concentration
were used to indicate the progress and extent of the smog formation.
After the maximum ozone concentration had been reached, the chamber
contents were evacuated through glass fiber filters. Several such runs
were made with each of the following hydrocarbons: a-pinene, cyclo-
hexene, 1-heptene, and toluene.
Preliminary analysis involved infrared spectroscopic examination
of particulate samples on millipore paper for use with a dual beam spectro-
photometer« Infrared spectroscopic analysis was also carried out on
organic extracts of particulate. Analysis revealed the presence of
various chemical classes and provided general information on similarities
and differences in particulate matter from various sources. Subsequent
analyses were of a more detailed nature and involved chemical fractiona-
tion of particulate extracts and identification of specific components
by gas chromatography (GC) and gas chromatography combined with mass
spectroscopy (GC-MS).
Analyses of urban aerosol and primary auto exhaust particulate
were directed toward the elucidation of differences between the organic
fractions of aerosol from the two sources. Specifically compounds were
sought which might be utilized as "fingerprints" of primary auto exhaust
to aid in establishing the automotive contribution to urban haze. Analyses
of the organic fraction of auto exhaust particulate by GC-MS revealed the
presence of benzoic acid and phenylacetic acid. These aromatic acids were
not detected in urban air particulate carried through the same analytical
procedure. The observations may be explained in terms of dilution of
auto exhaust in the atmosphere, or on the basis of removal of these acids
from the atmosphere by (secondary) photochemical reactions.
BATTIHLLE COLUMBUS
-------
A variety of other important differences between auto exhaust
and urban aerosol were determined. The concentration of aldehydes, ketones,
esters, and organic acids in urban aerosol is an order of magnitude greater
than that found in auto exhaust particulate. Presumably, the higher
concentration of oxygenates in urban aerosol results, at least in part,
from atmospheric reactions of hydrocarbons. Such a finding is consistent
with our studies of hydrocarbons reacted under simulated atmospheric
conditions. Two general types of reactions were observed, oxidative
cleavage of a carbon-carbon multiple bond, and oxidative decarboxylation.
Analyses of natural haze aerosol (Blue-Ridge Mountains site)
revealed a far lower concentration (~ 10 percent) of long chain carboxylic
acid than found in urban aerosol. This most likely reflects the presence
of urban sources contributing such organic acid pollulants or J:heir
(reduced) precursors to the atmosphere.
The smog chamber reactions of a-pinene typify the types of
reactions observed under simulated atmospheric conditions. This natural
terpene is produced in large quantities by plants and trees, and is found
in the forest atmosphere. Under conditions of the smog chamber it under-
goes oxidative cleavage to pinonic acid. Further reaction yields
"nor-pinonic" acid by oxidative decarboxylation. It is most significant
that analyses of natural haze aerosol revealed the presence of pinonic
acid. Pinonic acid was not found in urban aerosol carried through the
same analytical procedure. Ultra-low levels of this acid in urban aerosol
might be determined by ion-specific spectrometry. The level of pinonic
acid in urban atmosphere might be used as a measure of the contribution of
natural haze to urban smog where a forested region is contiguous to an
urban center. Finally, it should be noted, that the finding of pinonic
acid in both natural haze and smog chamber reaction products tends to
indicate that the smog chamber operating parameters used in this study
generate oxidizing conditions.
Details concerning the elucidation of reactions observed under
simulated atmospheric conditions are presented in the report section on
BATTELLE COLUMBUS
-------
Analysis of Smog Chamber Aerosol. An especially noteworthy result of this
phase of the research, however, was the finding of an organic nitrate
reaction product of cyclohexene. This finding reveals another sink for
nitrogen oxides and should help in establishing a better nitrogen balance.
EXPERIMENTAL
SAMPLING SITES
Field sampling of haze particles was performed at two locations
in New York City and in the Blue Ridge Mountains. Samples were collected
at the Bronx State Hospital grounds during the full month of September, 1970.
From September 21, through October 2, 1971, samples were also collected in
mid-Manhattan on the rooftop (fifth floor) of the Cooper-Union Building.
Battelle's mobile laboratory was then moved to the Blue Ridge Mountains
where continuous sampling was conducted from October 12, to October 23,
1970.
Bronx Site
The originally scheduled sampling site was in the Bronx, New York
City, on the grounds of the Bronx State Hospital. This location was chosen
because Scott Research Labs were conducting a gas phase monitoring program
at the same location for the Coordinating Research Council. The site
selection was based on the fact that the Bureau of Abatement, Environmental
Protection Agency, had previously monitored at that site.
This site proved to be less than adequate for aerosol studies.
The site was near the eastern edge of the Bronx, closa to Pelham Bay. As
a result, in the morning the air was generally clear due to a shore breeze
from the Bay. In the afternoon the wind would change and aged pollutants
from other areas of New York City would be convected to the site. The
trailer was located adjacent to a sewage pumping and chlorination station
ATTELLE COLUMBUS
-------
on the hospital grounds. There were apparently occasional leaks which
contributed to nighttime and unexpectedly high oxidant readings. The site
was located across the street from an area used for storage and repair of
trains used in the New York subway system. Within a three block radius
there was an elevated line, an expressway, two power plants for the State
Hospitalt a bakery, and a bulk oil and coal depot.
Due to instrument difficulties and scheduling problems the Scott
Research Lab monitoring effort was unable to obtain good measurements of
CO until well into the month of September. Since this was a critical
parameter for relating aerosol concentrations to auto exhaust emissions,
it does not seem useful to list in detail other measurements since they
are valueless without the CO data.
Due to inadequacies of the site both in terms of continuous
aerosol monitoring and in the unrepresentative nature of the collected
aerosol, it was decided to move the monitoring instruments to a site in
downtown Manhatten. However, hi-vol samples were collected continuously
at the Bronx site during the month of September.
Cooper Union Site
Arrangements were made through the New York City Air Pollution
Agency, in cooperation with Cooper Union, to set up a monitoring station
on the fifth floor roof of the Engineering Building of Cooper Union at
51 Astor Place. This station was adjacent to the New York City Air Pollu-
tion Agency's offices and laboratory. Integrating nephelometer measure-
ments and total particulate collections were made at this site during the
last two weeks of September. The building was located^at the intersection
of three major traffic arteries. However, due to the height there should
have been good mixing of auto exhaust. Two power plants were visible from
the roof but they were not close enough to make a predominant contribution
to the particulate matter collection. There did not appear to be any
other strong sources in the immediate neighborhood.
BATTEULE COLUMBUS
-------
Blue Ridge Mountain Site
At the end of September the sampling instruments were put back
into the trailer and the trailer was moved to a site in the Blue Ridge
Mountains in North Carolina. The site was located in the Pisgah National
Forest about 30 miles northeast of Ashville, North Carolina. It was
located between the two ranges of the Appalachian Mountains approximately
5 miles east of Mount Mitchell and 2 miles northwest of the Blue Ridge
Parkway at an altitude of 3600 ft. There was extensive evergreen and
hardwood forest cover. The trailer was located 50 feet from a gravel
road. Traffic consisted of only a few cars a day. The site was one
mile from the Black Mountain Campground but except on weekends there were
only 1 or 2 campers. (The site originally planned for was 5 miles from
the campground but this turned out to be on a hill below the gravel road
situated so that the dust cloud from the road was at the same height and
less than 25 feet from the intake ducts. It was therefore necessary to
relocate after observing the original site.) There were only a few houses
within a 5-mile radius of the site. The nearest community was Micaville
which was 15 miles away. Mount Mitchell and the mountain ranges protected
the site from pollution from any of the urban centers which would have
caused difficulties had the site been located outside the two ranges.
COLLECTION AND GENERATION OF AEROSOL
The collections of air-borne particles were made with two hi-vol
samplers, one conventionally designed and one equipped with a cyclone which
"cut out" nonrespirable and nonlight-scattering particles (> 2 microns).
The performance characteristics of the cyclone are presented under Cyclone
Design and Performance Evaluation, pages 9 - 14. Samples were obtained
on both glass and nylon fiber filters.
At all three sites, simultaneous measurement of visibility
reduction was made with an integrating nephelometer. Concurrent measure-
ments of the gas phase components were expected to have been made at the
BATTELLE COLUMBUS
-------
8
the Bronx site by Scott Laboratories, but scheduling precluded the
collection of most of that data. Monitoring in the Blue Ridge Mountains
included sunlight intensity, relative humidity, ozone concentration,
condensation nuclei count, and light scattering with an integrating
nephelometer.
Samples of light-scattering particles produced under controlled
conditions in the laboratory were collected with similar facilities
(hi-vol samplers). The laboratory aerosols were generated from two un-
ambiguous sources, primary automobile exhaust and irradiated N0~-hydrocarbon
mixtures.
The engine-exhaust aerosol was generated from a 1967 Chevrolet
automobile operated on a chassis dynamometer. The exhaust from one 7-mode
Federal driving cycle was vented into a large (300-cu-ft) Mylar bag con-
taining sufficient dry nitrogen to prevent condensation of water vapor.
The engine aerosol (primary) was collected by evacuating the bag through
a glass-fiber filter.
The aerosols generated from individual hydrocarbons were pre-
pared in a 610-cu-ft environmental chamber by irradiating 10 ppm (v/v)
selected hydrocarbon and 2 ppm of NO in clean, humidified air.
Measurements of visibility reduction and oxidant concentration
were used to indicate the progress and extent of the smog formation.
After the maximum aerosol concentration had been reached, the chamber con-
tents were evacuated through glass fiber filters. Several such runs were
made with each of the following hydrocarbons: Q/-pinene, cyclohexene,
1-heptene, and toluene.
BATTELLE COLUMBUS
-------
ORGANIC ANALYTICAL PROCEDURES
The organic fraction of the particulate was obtained by Soxhlet
extraction with "distilled-in-glass" methylene chloride for 6 hours.
Approximately 20 ml of solvent was used for each 4-inch filter disc.
Typically, the methylene chloride was distilled off and the residue
weighed. The "organic soluble residue" thus obtained was used for sub-
sequent fractionation and analysis.
Gas chromatography interfaced with mass spectrometry (GC-MS) was
frequently employed in the analysis of materials isolated from particulate
matter or smog chamber aerosol The instrument employed was a Finnigan
1015 Quadrupole Mass Spectrometer,, The ion source of this instrument has
been modified to permit the generation of either electron impact or
chemical ionization mass spectra. Instrument control and data acquisition
are accomplished using an interfaced digital computer. The reconstructed
gas chromatograms shown in this report are computer generated. The
detector response in such chromatograms is derived from integrated total
ion-current obtained during successive scans on the mass spectrometer.
The complete mass spectrum obtained during any given scan can be recalled
from memory (tape) by reference to a Spectrum Number along the x-axis of
the reconstructed gas chromatogram.
RESULTS AND DISCUSSION
CYCLONE DESIGN AND PERFORMANCE EVALUATION
A simple cyclone having a tangential air-flow inlet was designed
and evaluated for use as a precollector for large particles and for use
upstream of the filter. The intent of the cyclone was to remove atmospheric
particles that, because of their large size, were not effective in
(1) Obtained from Burdic and Jackson, Muskegon, Michigan.
BATTELLE COLUMBUS
-------
10
contributing to visibility reduction or the optical properties of haze.
This upper limit for haze particles was arbitrarily chosen to be about
2 Jim. Hence, the cyclone was intended to remove particles larger than
2 nm at the flow rate of the sampling system.
Because the particle-collecting ability of a cyclone is
dependent on volumetric gas flow rate, the cyclone design was checked
analytically to be sure that the total system pressure loss (pressure
loss in the cyclone plus loss in the filter) was not so large that the
capabilities of the blower were exceeded.
The design procedure for a simple cyclone having a tangential
air inlet was taken from the standard design equations as given in the
(2)
Chemical Engineers Handbook. A sketch of the general design of the
cyclone and the relationships between the dimensions are given in
Figure 1. Because the dimensions are all specifically related,the design
is based on only one dimension from which the others are then scaled.
(2)
The characteristic cut-off size, D , for the cyclone is given by:
D
B
c
>J£T1"e^(p«~p)
PC x Q -i-p
(D
where;
B = width of inlet duct, cm
N = factor depending on inlet geometry and flow rate,
approximately equal to 5
V = inlet gas velocity, cm/sec
C 3
p = density of particle material, gm/cm
3
p = density of gas, gm/cm
D = diameter of particle defined as characteristic
pc
cut-off size, cm
|j, = viscosity of gas, poise.
(2) Perry, J. H., Editor, Chemical Engineers Handbook. 3rd Edition,
McGraw-Hill Book Company, Inc., New York (1950), pp 1024-1026,
BATTELLE COLUMBUS
-------
11
Dc/4
oe/2
De/2
20e
DC/a
2 De
arbitrary,
tonally V. ft
Section A-A
Dust \ out
FIGURE 1. CYCLONE SEPARATOR PROPORTIONS
BATTELLE COLUMBUS
-------
12
The characteristic cut-off size is defined as the diameter of those
particles collected by the cyclone with an efficiency of 50 percent.
The pressure loss in a cyclone of this design is given as:
AP = FCV &. 0030 p* (V*)^j (2)
where:
AP = pressure loss, inches of water
3
p* = gas density, Ib/ft
V* = inlet velocity, ft/sec
F = constant characteristic of cyclone geometry which
for this design is reported to have a value of 8.
An additional consideration is the volumetric flow rate desired
3
for sampling. In this case 20 ft /min was chosen as being reasonable
from considerations of sampling time, pressure loss, and capacity of the
high-volume sampler blower. The inlet velocity to the cyclone, V , is
given in terms of volumetric flow rate, Q, and cyclone dimensions as:
V = Q/2B 2. (3)
c c
Equations (1} and (3) can then be combined to give
Tl =
c
2
e
-------
13
Because the cyclone design procedure may not be accurate for
small particles, it was felt that the design should be evaluated after
the cyclone was constructed. Calibrations were performed with two test
aerosols, dyed water droplets and a selected portion of small, dry dust
particles separated from A.C. Test Dust, Fine.
The water aerosol with uranine dye tracer added was produced
with a DeVilbiss D-40 nebulizer operating with the side port closed and
at an air pressure of 5 psig. The size distribution of the resulting
aerosol was measured using fluorometric techniques and a Battelle cascade
impactor. Three determinations of the size distribution were made and
are in excellent agreement as shown in Figure 2.
A.C. Test Dust, Fine was separated in a BAHCO particle classifier
and those particles separated as being smaller than 10.4 |j,m were used to
form the dust aerosol. This fraction of small particles was then aspirated
into an airstream for use as a test aerosol. The size distribution of
dry dust particles used in this case was determined with a Coulter Counter
and this distribution is shown in Figure 2.
The efficiency of the cyclone was determined for each test aerosol
by analyzing, either fluorometrically or by weight, for the amount collected
on a glass fiber filter downstream of the cyclone. The glass fiber filter
was MSA 1106B and should have an efficiency of essentially 100 percent for
the particle sizes of interest in this study.
Three determinations with the water plus dye aerosol gave cyclone
efficiencies of 82, 89, and 85 percent. Two determinations with the dry
dust gave efficiencies of 89 and 84 percent. The mass mean size for the
water plus dye aerosol was 2.4 |j,m and for the dust was about 5 ^m.
Although an exact determination of cut-off size was not made, it is
estimated from comparisons between the size distributions and cyclone
efficiencies that for the water aerosol the cut-off size, D , is about
1 p,m or slightly less. For the dry dust the cut-off size is about 2.5 |im.
The lower efficiency for the dry dust as compared with the water plus dye
aerosol is undoubtedly the result of reentrainment from the walls of the
BATTELLE COLUMBUS
-------
14
:-H HI-i
r=7--^-r-p{-r->-.y tl
A.C. Test Dust Fine (<10.4 ^m)
FIGURE 2. SIZE DISTRIBUTIONS OF TEST AEROSOLS
-------
15
cyclone for the dry dust. Reentrainment is known to become significant
for small diameter cyclones and apparently is significant enough to over-
come the effects of material density. In either case it appears that the
cyclone operated in approximately the predicted manner and gave cut-off
sizes near the design value of 2 (j,m.
LIGHT SCATTERING STUDIES
Light scattering studies were made at the three sampling sites
with an integrating nephelometer. Histographs are shown in Figure 3.
Average visibility during the periods studied was 40 miles for the Blue
Ridge Mountains site, 12 miles for the Bronx site, and 9 miles for the
Manhattan site. As shown on the histograph, visibility in the Blue Ridge
Mountains was never less than 14 miles. The blue haze is noticeable because
there are many vistas where mountain ranges can be seen 50 to 60 miles in
the distance. The lowest visibility recorded at the Bronx site was 5 miles
and at Cooper-Union, 3 miles.
Observations of terpenes by gas chromatography and of ozone by
Regner and Mast instruments tend to indicate that the blue haze is a photo-
chemical aerosol generated from terprene emissions from trees. Several
profiles showing the relationships of light scattering, sunlight, and
ozone observed in the Blue Ridge Mountains are shown in Figure 4. Two
light scattering profiles from the Cooper-Union site are shown in Figure 5.
These illustrate the effect of different meteorological conditions.
BATTELLE COLUMBUS
-------
16
8
10
Integrating Nephelometer, b , in 10" nf
SC3 u
30 15 10 7.5 6 5 4
Visibility, miles
FIGURE 3. FREQUENCY DISTRIBUTION OF VALUES OF VISUAL
RANGE MEASURED BY AN INTEGRATING NEPHELOMETER
AT THREE SAMPLING SITES: BRONX, N.Y., MID-
MANHATTAN, N.Y. (COOPER-UNION BIDG), AND
BLUE RIDGE MOUNTAINS
-------
17
0600 0300
/OOO
/ZOO
1400 /600
MOO
FIGURE 4. DAYTIME-CONCENTRATION PROFILE OF THE SUNLIGHT
INTENSITY, OZONE, AND LIGHT-SCATTERING
RELATIONSHIPS OBSERVED IN THE BLUE RIDGE
MOUNTAINS.
-------
18
a 14
12
10
<0
JC 6
CL
g> 4
rl
4J
!H 2
- 2.5
- 3
CO
0)
-3.75 £
^-i
1-1
XI
i-i
-5 «
-7.5
.15
9/25/70
hot, clear, calm
9/29/70, cold, cloudy, windy
i
5
678
Time of Day, a.m.
FIGURE 5. EFFECT OF METEOROLOGICAL CONDITIONS ON VISIBILITY
IN MANHATTAN (COOPER-UNION SITE)
-------
19
ORGANIC ANALYTICAL STUDIES
Infrared Spectroscopic Analysis of Total Particulate
The use of infrared spectra to give information about the
organic content of aerosols was investigated. Particulate samples
associated with urban and rural atmospheric aerosol and auto exhaust
were examined. Aerosols generated in the smog chamber from straight
chain and monocyclic olefins and aromatics were also examined. (Earlier
/ *3 / \
studies ' describe the formation of aerosol from cyclic and bicyclic
olefins.) Infrared spectra obtained from aerosol particles collected
directly on Millipore filter paper (ultrathin type TH, 0.45 pm) are
shown in Figures 6 through 10.
Spectra of aerosols generated from cyclohexene/NOjj and m-xylene/NQjj
are shown in Figure 6. Absorption bands due to the OH stretch in organic
acids, the CH stretch, and the carbonyl C=0 stretch can be identified. The
low frequency part of the doublet at about 6 microns is due either to C=C
or organic nitrates. The similarity between the cyclohexene and the
m-xylene spectra is evident. In addition to the organic absorption bands,
other features may be attributed to NIL and possibly S0^~ and NO^ .
In Figure 7 spectra are compared from 1-heptene/NQjj and from
l-heptene/NOjj/802. Several new peaks can be seen in the spectrum of the
aerosol from the 862 containing system. The infrared spectra of the
methylene chloride extracts however are identical, indicating that the
additional material in the SCL-containing aerosol is possibly inorganic.
A comparison of the spectra of primary and secondary aerosols
from auto exhaust is shown in Figure 8. The relative reduction of CH
(3) Groblicki, P. J., and Nebel, G. J., "The Photochemical Formation of
Aerosols in Urban Atmospheres", General Motors Corporation Research
Publication GMR-957, October 6-7, 1969.
(4) Endow, H., Doyle, G. J., and Jones, J. L., "The Nature of Some Model
Photochemical Aerosols, JAPCA, .13, 141 (1963).
BATTELLE COLUMBUS
-------
Wavelength, microns
2.5
3.5
4.5 5
i i i i i
ro
o
FIGURE 6. INFRARED SPECTRA OF AEROSOLS GENERATED FROM CYCLOHEXENE/NO AND M-XYLENE/NO
x x
-------
21
4000
0.2O
3OOO
2500
Wovcnumbei, cm"1
2000 I6OO 1300
1000
800
5678
Wavelength, microns
10
3000
2500
Wavenumber, cm"
2000 1500 1300
IOOO
800
I
1
0.10
OJO
-------
Wavelength, microns
567
Primary Auto-Exhaust Aerosol
Secondary (Irradiated) Auto-Exhaust Aerosol
to
tv>
FIGURE 8. INFRARED SPECTRA OF PRIMARY AND SECONDARY AUTCH EXHAUST AEROSOL
-------
23
Y
300O
2500
1800
Wavenuraber, cm
-1
/too
FIGURE 9. COMPARATIVE INFRARED ABSORPTION BANDS OF AEROSOL
GENERATED FROM KNOWN PRECURSORS (1-HEPTENE-NO )
AND FROM COMPLEX MIXTURES OF URBAN SOURCES X
-------
24
BWf WDGC MOUNTAIN A&09QI.,
W
-I \r
3100
500O
IBOO
-1
. 1600
/400
/200
Wavenumber, cm
FIGURE 10, COMPARATIVE INFRARED ABSORPTION BANDS OF AEROSOLS
GENERATED IN A SMOG CHAMBER FROM A MIXTURE OF A
PURE TERPENE (a-PINENE) WITH NOX AND IN A NATURAL
FORESTED ATMOSPHERE (BLUE RIDGE MOUNTAINS)
-------
25
absorption, and the increase in organic acid OH, C=0, and NC^ or C=0 ab-
sorption in the photochemical aerosol can be seen. Unfortunately the
resolution in the spectra of samples held on filters is low and there do
not appear to be any characteristic absorptions that can be used to dif-
ferentiate primary and secondary aerosols from automobiles.
In Figures 9 and 10 are shown portions of the infrared spectra
obtained by extracting the aerosol material with methylene chloride and
evaporating the solution to dryness on a sodium chloride plate. Consid-
erably better resolution is obtained in the 6-micron region. The.C=0
peaks have been observed and commented on previously but the identifi-
cation of the 6.2-micron peak with C=C and organic nitrates is believed
to be new. This assignment has been supported by the observation, with
gas chromatography-mass spectroscopic techniques, of unsaturated organic
acids in aerosols formed from 1-heptene/NO and a nitrate ester in the
X
aerosol.1; formed from cyclohexene/NOx.
In Figure 9, the spectra of aerosols from 1-heptene/NC) and
X
from atmospheric aerosols collected at the Cooper-Union site are compared.
In Figure 10, a similar comparison from a-pinene/NOx, and the Blue Ridge
Mountains aerosol is shown. The similarities are quite evident, the major
differences being in the height of the nitrate or C=C peaks. The series
of absorptions between 8-12 microns have not been identified.
Analysis of Particulate Extracts;
Sample Sources and Analytical Objectives
Organic extracts of particulate from the following sources were
prepared:
(1) Urban and rural atmospheric haze
(2) Primary auto exhaust
(3) Aerosol generated from known precursors in
Battelle-Columbus' smog chamber.
(5) Renzetti, N. A., and Doyle, G. J., "The Chemical Nature of the
Particulate in Irradiated Auto Exhaust", JAPCA, .8, 293 (1959).
ATTELLE COLUMBUS
-------
26
Urban air particulate was collected at the Cooper-Union sampling
site during the periods 9/23/70 through 9/26/70 and 9/29/70 through 10/1/70.
The 4-day period beginning 9/23/70 was extremely hot and clear and was
marked by fairly intensive photochemical smog episodes as evidenced by the
oxidant levels shown in Figure 11. For comparison, the maximum hourly
average total oxidant measured at this site between July 20 and October 1
exceeded 10 pphm on only 3 days (measurements made by the New York Air
Resources Laboratory). Following two days of rain, the period beginning
9/29/70 was quite cool and cloudy and very little photochemical smog was
evident. Visibility during the collection periods is indicated in Figure 1.
A conventional hi-vol sampler was used to collect aerosols at the Cooper-
Union site. Facilities were not available there to measure total sampling
size or total particle mass.
Rural-air particulate was collected with both a 2-stage cyclone
sampler and a conventional hi-vol sampler. The particulate analyzed was
collected exclusively with the eyelone-equipped sampler during clear
weather in the period 1/18/70 through 10/23/70. On one day in that period
(10/20/70) it rained. During clear weather a blue-grey haze was noticeable
in the daytime. The air flow rate through the conventional sampler aver-
aged 16 cfm; that through the cyclone sampler averaged 14 cfm. The par-
3
ticulate mass concentration averaged 31 u>g/m with the conventional
3
sampler; 26 M-g/m with the cyclone sampler. On this basis it appears
that the diameters of aerosols associated with forested areas are predomi-
nantly in the light-scattering size range or smaller.
Primary auto-exhaust particulate was generated from a 1967 auto-
mobile operating on a chassis dynamometer. The method of sample collection
was described earlier.
Initial analytical studies were concerned with defining general
similarities and differences between urban air particulate and primary
auto exhaust particulate. Particulate matter from the two sources was
fractionated into organic acid, basic, and neutral classes; the general
distribution of material with respect to organic chemical class was thus
determined. Subsequent work involved infrared spectroscopic, gas chroma-
tographic, and mass spectral analysis of various components of the
AT T E U L E COLUMBUS
-------
27
10
8
t 6
Q.
9/23/70
9/24/70
I i I
II
II
10
8
e
f i. e
CL
9/25/70
9/26/70
II
FIGURE 11. HOURLY AVERAGE PLOTS OF TOTAL OXIDANTS,
COPPER-UNION SITE, NEW YORK CITY
BATTELLE COLUMBUS
-------
28
particulate matter. Specifically, auto exhaust components were sought
that could be utilized as "fingerprint" compounds to establish the contri-
bution of auto exhaust to atmospheric pollution.
Several components of aerosol produced in the smog chamber from
known precursors were identified by mass spectral techniques. These analy-
ses of smog chamber reaction products were conducted with the objective of
elucidating the nature of reactions occurring under simulated atmospheric
conditions.
High-pressure liquid chromatography (HPLC) was evaluated with
respect to its effectiveness in the separation and assay of polynuclear
aromatic hydrocarbons (PNA's). This technique is especially suited to the
analysis of compounds whose low volatility makes gas chromatographic sepa-
ration difficult. Further, HPLC is compatible with the use of ultraviolet
absorbence and fluorescence techniques for peak detection. The utility
of HPLC in the chromatographic separation of PNA's was demonstrated.
Comparison of Air Particulate
and Auto Exhaust Particulate
Analysis of urban air particulate and primary auto exhaust was
undertaken with the general objective of elucidating similarities and dif-
ferences in the organic fraction of material obtained from these sources.
Initial analyses determined the distribution of material among organic
acid, basic and neutral fractions. The above fractions were analyzed by
infrared spectroscopy to determine the presence and approximate concentra-
tions of compounds bearing specific functional groups. Subsequent study
of both urban and rural air particulate, as well as auto exhaust particu-
late, involved detailed analysis of the acid fraction of these materials
using the techniques of thin-layer chromatography, gas chromatography, and
^
gas chromatography interfaced with mass spectroscopy (GC-MS). It was the
objective of this phase of the work to identify specific components of auto
exhaust that might be utilized as "fingerprint" compounds to establish the
contribution of auto exhaust to atmospheric pollution.
BATTELLE COLUMBUS
-------
29
Chemical Fractionation and Infrared
Spectroscopic Analysis
The methylene chloride extracts of urban air particulate and
auto exhaust particulate were fractionated according to the scheme shown
in Figure 12. As indicated, the residues obtained upon removal of methyl-
ene chloride by distillation were extracted with redistilled diethyl ether;
79 percent of the air particulate residue went into ether solution, while
85 percent of the auto exhaust residue went into ether solution. Infra-
red spectroscopy indicates that the aromatic hydrocarbon content is
severalfold higher in the ether insoluble fraction than in the ether soluble
fraction.
The approximate material distribution among the organic acid,
basic, and neutral fractions of air and auto exhaust particulate is shown
below.
Material Distribution
Acid Basic Neutral
Fraction Fraction Fraction
Air particulate ~ 24%, ~ 38%, ~ 3870
Auto exhaust ~ 26% ~ 16% ~ 58%
particulate
Infrared spectroscopy of the neutral fraction of the above
samples reveals the presence of significant concentrations of ester,
aldehyde, and ketone in material isolated from air particulate. The
ester concentration is greater than that of the combined aldehydes
and ketones; the concentrations are in an approximate ratio of
r-'
(ester):(aldehyde + ketone) = 10:1. Such carbonyl and carboxyl contain-
ing compounds are largely absent (i.e., < 5%) from the neutral fraction
of auto exhaust particulate. The neutral fraction of auto exhaust parti-
culate is almost entirely hydrocarbon, with an (aliphatic):(aromatic)
ratio of approximately 85:15. This is approximately the ratio observed
in the hydrocarbon component of urban air particulate.
BATTELLE COLUMBUS
-------
30
Methylene Chloride
Extract of
Particulate Matter
Distill Methylene Chloride
Estract with Diethyl Ether
Ether Solution
Dissolve in
Methylene Chloride
Methylene Chloride
Solution
Extract with 1 N HC1
. Ether Solution
Neutralize,
Make Basic
Aqueous Phase
Containing
Organic
Ba.se Salts
Extract with 2 N NaOH
Neutralize,
Make Acidic
Ether
Solution
of Organic
Neutrals
Aqueous
Phase
Extract with Diethyl Ether
Ether Solution
of Organic
Acids
FIGURE 12. SCHEME FOR ORGANIC CLASS-SEPARATION
-------
31
Isolation of the basic fraction of urban air particulate was
attempted by extraction of the ether soluble fraction with IN HCL. The
HCl extract was neutralized, made basic (pH = 13), and extracted with
diethyl ether to recover the organic base. Most of the basic material,
however,, was not recoverable by ether extraction of the aqueous phase.
Approximately 38 percent of the total ether soluble fraction went into
solution upon extraction with 1 N HCl. However, when the HCl extract
was made basic only ~2 percent of the ether soluble material could be
recovered. Note in addition that previous extraction of the total ether
soluble fraction with distilled water did not indicate the presence of
an appreciable water soluble component.
The solubility behavior of this relatively hydrophillic com-
ponent of the basic fraction tends to indicate that the compounds present
are polar in nature and are polyfunctional. The presence of polar func-
tional groups would tend to increase the hydrophillic character of the
molecule, rendering recovery of the material by extraction from an aqueous
phase more difficult.
Infrared spectroscopy of the ether extractable base revealed
the absence of primary, secondary, or tertiary amines. The infrared data
tend to indicate the presence of nitrogen containing heterocyclic aroma tics
such as pyridine or acridine.
Infrared spectroscopy of the acid components of the above samples
indicates a greater concentration of alcoholic hydroxy containing compounds
(R-OH and ArOH) in the fraction isolated from auto exhaust particulate than
that isolated from urban air particulate. That is, the ratio of (COH):
(C-H) is several fold higher for auto-exhaust particulate than urban air
particulate. The (aromatic):(aliphatic) ratio is approximately 5:95 for
air particulate and approximately 75:25 for auto-exhaust particulate.
Analysis by Gas Chromatography and Gas Chromatography
Interfaced With Mass Spectrometrv
Organic acids isolated from urban and rural air particulate and
from primary auto exhaust particulate were analyzed using gas chroma-
tography (GC) and gas chromatography interfaced with mass spectroscopy
BATTELLE COUUMBUS
-------
32
(GC-MS).' Methyl esters of the acids were prepared by reaction with
diazomethane. The products were gas chromatographed, using a 10-ft x
1/8-in. column of 10 percent ethylene glycol succinate (EGS) on Gas
Chrom Q, as well as a 10-ft x 1/8-in. column of 3 percent OV-17 on Gas
Chrom Q. GC analysis of esterified urban air particulate acids re-
vealed the presence of various long-chain carboxylic acid esters,
principally palmitate and stearate with lesser concentrations of laurate,
myristate and arachidate. Initial peak assignments were made on the
basis of comparative retention times using authentic samples of the
various methyl esters. The presence of methyl palmitate and methyl
stearate was confirmed by GC-MS. Several other tentatively identified
ester peaks did not produce discernible parent-ion peaks (GS-MS), but
did produce fragmentation patterns characteristic of carboxylic acid
methyl esters. Similar gas chromatographic analysis of material isolated
from auto-exhaust particulate revealed a far lower concentration (~10%)
of long-chain carboxylic acid esters than that observed in urban air
particulate. On the basis of gas chromatographic retention times,
lauric, myriatic, palmitic, and stearic acid esters can be tentatively
identified. Gas chromatographic analysis of acids isolated from rural
air particulate associated with natural haze similarly revealed a
carboxylic acid concentration lower than that found in urban air par-
ticulate. Lauric, palmitic, and stearic acid esters can be tentatively
identified on the basis of gas chromatographic retention times. Sub-
sequent analysis of natural haze acids by GC-MS confirmed the presence
of methyl palmitate and also revealed the presence of methyl pinonate.
Pinonic acid is an oxidation product of orpinene, a natural terpene
present in forest atmosphere. This finding is discussed in detail in
the report section dealing with analysis of smog-chamber aerosol gen-
erated from a-pinene.
Comparison of the above gas chromatograms indicated the pre-
sence of materials in the auto-exhaust particulate which were absent
(or present in negligible concentration) in air particulate. To sim-
plify gas chromatographic analysis of the complex mixtures of acids,
the esterified samples were fractionated by preparative thin-layer
BATTELLE COLUMBUS
-------
33
chromatQgraphy (TLC) on silica gel. Plates were developed with chloroform
and visualized using ultraviolet illumination. The various TLC fractions
were analyzed gas chromatographically. A principal difference between
auto exhaust and air particulate was noted in the gas chromatogram of TLC
fraction No. r (rt = 0.35 - 0.70). Two major peaks in the chromatogram
of auto-exhaust material were absent in that of air-particulate material
(see Figures 13 and 14). Mass spectroscopy was performed to characterize
the two compounds in question. The mass spectra are unambiguous and es-
tablish the two peaks as corresponding to methyl benzoate (Peak A) and
methyl phenylacetate (Peak B). The mass spectra are shown in Figures 15
and 16.
Compounds such as the two identified above could possibly be
utilized as "fingerprint" compounds to establish the contribution of pri-
mary auto exhaust to atmospheric pollution. Air- and auto-exhaust par-
ticulate samples from a variety of other sources should be analyzed to
confirm the feasibility and reliability of such "fingerprint" utilization
in procedures to establish automotive contribution to atmospheric pollution.
Analysis of Smog Chamber Aerosol
A variety of pure compounds were reacted in Battelle's smog
chamber. Concentrations of reactants employed, the smog chamber operat-
ing conditions, and smog parameters are described in Table 1. Analyses
of reaction products generated were conducted with the objective of eluci-
dating the nature of reactions occurring under such simulated atmospheric
conditions. In view of the oxidizing conditions that obtain under such
conditions, the acid (most oxidized) fraction of reaction products was
isolated for analysis by gas chromatography and mass spectroscopy.
Typically, the acid fraction was isolated by ^extraction of the
total methylene chloride extract with 2N NaOH. The basic aqueous extract
was then reacidified to pH ~ 1.5 and extracted with methylene chloride.
The methylene chloride solution was dried over anhydrous MgSO, and the
solvent was removed by distillation. A portion of the isolated acid was
esterified with diazomethane and the methyl esters produced were analyzed
using gas chromatography and gas chromatography interfaced with mass
spectroscopy (GC-MS).
BATTELLE COLUMBUS
-------
Methyl esters of acid fraction of air particulates
Column: 10' x 1/8", SS, 3% OV-17 on 100/120 Gas Chrom A
Programmed: 100 C-200 C at 2 C/mln
Base Attenuation: 1 x 10"
154 57 fro"
FIGURE 13. GAS CHROMATOGRAM OF METHYL ESTERS OF ACID
FRACTION ISOLATED FROM AIR PARTICULATE
-------
Major
Peak A
Major
Peak B
TLC Cut #4 of Auto Exhaust Methyl Esters
Column: 10' x 1/8" SS, 3% OV-17 on 100/120 Gas Chrom Q
Programmed: 100 C-2-0 C at 2C/mtn
Base Attenuation: 1 x 10
3 6 9 12 15 18 2I~ 24 2V 30 33 35 39 42 45 48 51 5!4 5*7 60
(minutes)
FIGURE 14. GAS CHROMATOGRAM OF TLC FRACTIONATED METHYL ESTERS OF
. . ACID FRACTION ISOLATED FROM AUTO EXHAUST PARTICUIATE
-------
36
SYSTEM 150 IS ON
SELECT MODE: OUTP
PRINT?:
PLOT?: Y
FILE NAME: 13-2
SPECTRUM NUMBER: 28
AMPLITUDE EXPANSION?:
MINIMUM VALUE %:
SUBTRACT BACKGROUND?: Y
BACKGROUND FILE NAME: 13-2
SPECTRUM NUMBER: 20
BACKGROUND AMPLIFICATION:
NORMALIZE ON:
SPECTRM NlfBER 28
g
6C-MS CF fUTO EXHFUST SfWLE NO- 1
_
_
1
1 I'l
40 SO 0
U * p»
n/ t
Llli i
70 60 SO 100 110
methyl phenylacetate
u
120 130 110 ISO 160 170
FIGURE 15. MASS SPECTRUM OF MAJOR PEAK B IN GAS
CHROMATOGRAM OF AUTO EXHAUST METHYL ESTERS
-------
37
SYSTEM 150 IS ON
SELECT MODE: OUTP,
PRINT?:
PLOT?: Y
FILE NAME: 13-2
SPECTRUft NUMBER: 10
AMPLITUDE EXPANSION?:
MINIMUM VALUE %:
SUBTRACT BACKGROUND?: Y
BACKGROUND FILE NAME: 13-2
SPECTRUM NUMBER: 5
BACKGROUND AMPLIFICATION:
NORMALIZE ON:
SPECTRUM NLfBER 10
g
8.
OF am) BttFUST
NO. 4
El.
*f
^t
COOCH
methyl benzoate
Jlii
4'IUr.U, ujnl,....,....!.
TTTT"l""'"T'l
I S3 0
E
rriTii|iii i irpmiitiT TIIIIDITJTIII IIIT|IT₯TTII
73 03 33 133 118 123 130
1S3 1G0 178 183
FIGURE 16. MASS SPECTRUM OF MAJOR PEAK A IN GAS
CHROMATOGRAM OF AUTO EXHAUST METHYL ESTERS
-------
38
TABLE 1. SUMMARY OF PURE HYDROCARBON RUNS IN
610-CU-FT SMOG CHAMBER
Hydrocarbon
Initial Conditions
Temperature, F
RH, percent
uy _,._ v" /
no , ppm
NO, ppm
N02 , ppm
Smog Parameters
Light scattering
-4 -1
b-scat max, 10 m
t-max, min
Oxidant (Mast meter)
C-max, ppm as 0,
t , min
max
a-Pinene
88
41
10
1
1
190
10
0.25
8
Cyclohexene
90
39
10
0
2
214
18
0.32
15
1-Heptene
90
41
10
1
1
126
80
0.79
75
Toluene
89
43
10
1
1
73
150
0.60
115
(a) All ppm units are vol/vol.
(b) Aerosol collections made after light-scattering reached a maximum.
ATTELLE C O I-U IVI B U S
-------
39
Aerosol .Generated from Cyclohexene
The aerosol generated from cyclohexene was collected and the
organic acid isolated according to procedures described above. A portion
of the isolated acid was treated with diazomethane and the resulting
methyl esters were analyzed by GC-MS. The reconstructed gas chromatogram
is shown in Figure 17. Several tentative identifications can be made on
the basis of mass spectra of the indicated chromatographic peaks.
(1) Major peak C can be tentatively
identified as the methyl ester
of the difunctional aldehydo-acid:
The mass spectrum is shown in Figure 18«
CS/O
(2) Major peak B can be tentatively
identified as the methyl ester
of 5-hydroxypentanoic acid:
The mass spectrum is shown in Figure 19.
(ID
(3) Major peak A can be tentatively
identified as the methyl ester
of the difunctional aldehydo-acid:
The mass spectrum is shown in Figure 20»
Ooo/r ('C^)
(III)
(4) Major peak D can be tentatively
identified as the nitrate ester
of the above methyl 5-hydroxypentenoate:
The mass spectrum is shown in Figure 210
Recall that the cyclohexene was reacted in
the smog chamber in the presence of NCL0
'00/7'
(IV)
BATTELLE - COLUMBUS
-------
SYSTEM 150 IS ON *°
.SELECT MODEI IFSS
CALIBRATE?t £
TITLE* GC-MS CGH4) OF CYCLOHEXJINE-N02 PRODUCTS
CALIBRATION FILE NAME! CAL-H A
FILE NAMEl 24-3
MASS RANGE) 50*350
SAMPLES/AMUt 1
MAX RPT COUNT* A
BASE INTEGRATION TIME! 8
RPT COUNT BEFORE CHECKING LOVER THRESHOLD) 4
LOWER THRESHOLD! 3
UPPER THRESHOLD! 4
1015 RANGE SETTING?! H
MAX RUN TIMEt 30
GC-MS CQW OF aCLGHEXa-N02 PRODUCTS
10 29 33 40
SPECTRM NUJ1SGR
TJTTr.prr.j
58 60
70 60 33
100 110 120 130
FIGURE 17. RECONSTRUCTED GAS CHROMATOGRAM OF
CYCLOHEXENE REACTION PRODUCTS
-------
SYSTEM 150 IS ON
SELECT MODEI OUTP
PRINT?:
PLOT?t Y
FILE NAME* 24-3
SPECTRUM NUMBER! 23
AMPLITUDE EXPANSION?:
MINIMUM VALUE SI
SUBTRACT BACKGROUND?t Y
BACKGROUND FILE NAMEI 24-3
SPECTRUM NUMBER! 20
BACKGROUND AMPLIFICATION*
NORMALIZE ON I
41
Tentative assignment:
COOCH,
.CHO
SPECTRLM HJCSSR 23
B
GC-MS CCWD OF CYCL0XWE-N02 PRODUCTS
o
D.
K
O
ttf
ftt
§
fl
-------
SYSTEM 150 IS ON
SELECT MODE! OUTP
PRINT?!
PLOT?* Y
FILE NAMES 24-3
SPECTRUM NUMBER! 18
AMPLITUDE EXPANSION?!
MINIMUM VALUE Si
SUBTRACT BACKGROUND?! Y
BACKGROUND FILE NAME! 24*3
SPECTRUM NUMBER* 15
BACKGROUND AMPLIFICATION!
NORMALIZE ON!
42
Tentative assignment:
CCOOCH3
CHjOH
SPECTRM NU-BSR 18
GC-ttS CCW) OF CYCUDHEXRC-W32 PRODUCTS
W *OM
8.
8.
ffi~
a.
$8-
bB.
U.ED
§
go
8.
2.
O-
10 SO 60 7C
iui v c*
fix t
tH
o
B
01
"B
1
1
JJ ,
J 80 30 1
If
-
1
'a
^N
f
+
^*
' %.
I,,!,,,.,, 4
00 11C
>
i
t
>
i*
i
X
eo
r-l
a ^
' at 1-1
i
*> a>
+i ^
^ + "
J *"
l,,,,,,,,,,,,!,,.!,,,..,!,,,!..,.,,.,,,,^,,!,,,,,,,,,,,,,,,,,,,,,,,,,,,,
J 120 130 140 1S0 160 170 180
FIGURE 19. ANALYSIS OF CYCLOHEXENE REACTION PRODUCTS:
CHEMICAL IONIZATION MASS SPECTRUM FOR GAS
CHROMATOGRAPHIC PEAK (B)
(Refer to Figure 17)
-------
SYSTEM 150 IS ON
SELECT MODEI OUTP
PRINT?I
PLOT?! Y
FILE NAMEJ 24-3
SPECTRUM NUMBER* 10
AMPLITUDE EXPANSION?!
MINIMUM VALUE J5«
SUBTRACT BACKGROUND?: Y
BACKGROUND FILE NAME: 24-3
SPECTRUM NUMBER* 6
BACKGROUND AMPLIFICATION*
NORMALIZE ONt
43
Tentative assignment:
CCOOCH
CHO.
SPECTRIM NU-BER 10
GC-MS COW OF
PRODUCTS
w » _
8.
F
bB.
ll
no
s
L
s"
O-
co
II
1
V
X
0
f
K
1
^
1 Ii 1 1 1
,1,,,,,..|.J.1).,M|....,....| ,,,. |.,,,,,
10 SO 60 70 80
H/ E
^~
p:
e
i
N_
........tl
M1"'1"1
3Q
i
co ii
i
OJ
I ^
en
ii
i
«-. m
B ^
S QJ
^
aT
CM
Ulllilllllllllll IlllllllllllltllllllllltlllllllllllllllllllH
100 110 123 130 HO ISO 160 170
FIGURE 20. ANALYSIS OF CYCLOHEXENE REACTION PRODUCTS:
CHEMICAL IONIZATION MASS SPECTRUM FOR GAS
CHROMATOGRAPHIC PEAK (A)
(Refer to Figure 17)
-------
SYSTEM 150 IS ON
SELECT MODS! OUTP
PRINT?t
PLOT?» V
FILE NAMES 24-3
SPECTRUM NUMBER! 39
AMPLITUDE EXPANSION?!
MINIMUM VALUE 7.1
SUBTRACT BACKGROUND?* Y
BACKGROUND FILE NAM2J 24-3
SPECTRUM NUMBER: 36
BACKGROUND AMPLIFICATIONS
NORMALIZE ONI
Tentative assignment:
CCOOCH3
CH2Om
SPECTRIM MttBEH 39
GC-MS CCW3 OF CYCLCHEXfiNE-hOZ PRODUCTS
O
§
0)
"B
00
V
V.
.. ,.,.,,-j-...,,.-..-,.-,--... 111 i , . . | _ j w | , ^ . j j f f ,
w so 60 70 ea so leo no 120 iaa 1^0 iso iee no iso is.
E
FIGURE 21. ANALYSIS OF CYCLOHEXENE REACTION PRODUCTS:
CHEMICAL IONIZATION MASS SPECTRUM FOR GAS
CHROMATOGRAPHIC PEAK (D)
(Refer to Figure 17)
-------
45
- Oxidative cleavage of cyclohexene in the presence of ozone and
^2°2 can reasonably be seen to yield compound (I). Decarboxylation
of (I) might yield a dialdehyde OHC-(CH ) -CHO which might then undergo
disproportionation of yield (II). Alternatively, Compound (I) might be
degraded at the aldehyde terminal to yield the hydroxypentanoic acid via
an unspecified free radical mechanism. Oxidation of the hydroxy terminal
of 5-hydrosypentaboic acid (II) yields the five membered aldehydo-acid,
(III).
In summary, several acidic reaction products of cyclohexene
have been isolated from smog chamber aerosol and tentatively identified.
Aerosol Generated from q-Pinene
a-pinene, a natural terpene, is produced in large quantities
by plants and trees. Significant concentrations of this fragrant
volatile compound can be found in the forest atmosphere. The aerosol
produced in the smog chamber from this compound was collected, and the
acid fraction was isolated as described above. Methyl esters were pre-
pared and analyzed by GC-MS. The reconstructed gas chromotogram is
shown in Figure 22. The mass spectrum of the indicated chromatographic
peak corresponding to methyl pinonate is shown in Figure 23. This peak
can be unequivocally identified as that corresponding to methyl pinonote
on the basis of comparison of its mass spectrum with that of authentic
methyl-cis-pinonote shown in Figure 24. The mass spectral fragmentation
pattern for this compound is rationalized in Table 2.
A second analysis of the sample by GC-MS verified the presence
of pinonic acid. The analytical results indicated, in addition, that
(6) B. W. Gay and J. J. Bufalini, "Hydrogen Peroxide in the Urban
Atmosphere", paper presented at 161st ACS National Meeting,
Los Angeles, March, 1971. Hydrogen peroxide was found in
irradiated mixtures of hydrocarbons and nitrogen oxides as
well as in urban atmosphere during periods of photochemical
smog formation.
BATTELLE COLUMBUS
-------
SYSTEM 150 IS ON 46
SELECT MOD1£: 1FSS
CALIBRATE?:
TITLE: GC-MS (CH4> OF METHYL ESTERS FROM PINENE REACTION
CALIBRATION FILE NAME: CAL-M
FILE NAME: 15-7
MASS RANGE: 50-250
SAMPLES/AMU: 1
MAX RPT COUNT: A
BASE INTEGRATION TIME: 8
RPT COUNT BEFORE CHECKING LOWER THRESHOLD: 2
LOWER THRESHOLD: 3
IPPER THRESHOLD: A
1015 RANGE SETTING?: ' M
MAX RUN TIME: 30
DATA
8
H
5L
BO-MS ccw) OF
ESTERS FEOM pita^: PJSFCTKN
methyl pinonate
0 10 23 S3 13 S3 0 70 £3 S3 1C3 110 133 1
FIGURE 22 . RECONSTRUCTED GAS CHROMATOGRAM OF
ct-PINENE REACTION PRODUCTS
(First Analysis)
-------
47
SELECT MODE: OUTP .
PK1NT?: '
PLOT?: Y
FILE NAME'S 15-7
SPECTRUM NUMBER: 96
AMPLITUDE EXPANSION?:
MINIMUM VALUE %:
SUBTRACT BACKGROUND?: Y
BACKGROUND FILE NAME: 15-7
'SPECTRUM NUMBER: '92
BACKGROUND AMPLIFICATION:
NORMALIZE ONJ
SPECTRUM
35
I HCTHVt ESTERS RTOM PIN5NS RSRCTICN
_ 90 .
80
70 1 '
w
(L(\ OT
60 a
... 50 w
o
40 8
- 30
- 20
- 10
o 1
.......1.1 l.ll
,,l,
(See Table 2 for explanation
of fragmentation pattern)
er>
V-l
a
i
j-,.^^ri.i;4ijiiiiiiiiiiviijviiij 1.1 |.i.r^ iv. .;:.. .'i.jlljllll. uil|lluijrill|J 11,1111 IJ
1S S8 63 78 63 S3 168 118 128 133 118 lErt 1C3 170 183 1S3 Z6S
FIGURE 23. ANALYSIS OF a-PINENE REACTION PRODUCTS:
CHEMICAL IONIZATION MASS SPECTRUM OF GAS
CHROMATOGRAPHIC PEAK CORRESPONDING TO
METHYL PINONATE
-------
SPECTRUM NUMBER 173
ec-MS cow OF METHYL ESTER OF CIS-PINQNIC ACID
100
- 90
, 80
- 70
- 60
-------
49
rv*i-,
,H
Protonated parent-ion of
methyl pinonate,
(MR)*, m/e" = 199
Fragmentation of Parent-Ion
Process
Loss of H20(MH -18)
Fragment
m/e
181
Loss of CH OH(MH+-32)
167
Loss of HCOOCH3(MH -60)
Transannular cleavage
139
111
Transannular cleavage
99
-Transannular cleavage
s»
"H
71
Transannular cleavage
129
TABLE 2. FRAGMENTATION OF METHYL PINONATE UNDER
CONDITIONS OF CHEMICAL IONIZATION
(METHANE) MASS SPECTROMETRY
BATTELLE COLUMBUS
-------
50
trans-pinonic acid may be present along with the more abundant cis isomer.
Interconversion of the isomers in the presence of ultraviolet irradiation
can be postulated. Tentative identification of the chromatographic peaks
corresponding to the isomeric pinonic acids is indicated on the recon-
structed gas chromatogram in Figure 25.
cis-pinonic acid
trans-pinonic acid
Tentative identification of the secondary reaction product of
a-pinene has been made. The compound corresponds to Peak A of the recon-
structed gas chromatogram in Figure 25. The mass spectrum of this compound
is shown in Figure 26. The compound is tentatively identified as the homo-
logous acid resulting from oxidative decarboxylation of pinionic acid, 2,
2-dimethyl-3-acetyl cyclobutanecarboxylic acid. The compound is herein-
after referred to as "nor-pinonic" aicd. Its mass spectral fragmentation
pattern is rationalized in Table 3.
Thus we see that under the simulated atmospheric conditions of
the smog chamber, Q/-pinene undergoes oxidative cleavage to pinonic acid.
The pinonic acid produced may then be oxidatively decarboxylated (tentative)
to the next lower homologous acid.
CH=
CH*
a-pinene
pinonic acid
"nor-pinonic" acid
BATTELLE COLUMBUS
-------
SYSTEM ISfc) IS ON
SELECT- MODE: CONT 51
CALIBRATE?:
TITLE: GC-MS (CH4) OF SAMPLE 28532
CALIBRATION FJLE NAME: CAL-H
FILE NAME: 27-2
MASS RANGE: 50-350
INTEGRATION TIME: 17
SAMPLES/AMU: I
THRESHOLD: 1
1015 RANGE SETTING?: H
MAX RUN TIME: 30
-45-28
SYSTEM 150 IS ON
SELECT MODE: OUTP
PRINT?:
PLOT?:
RECONSTRUCTED GAS CHROMATOGRAM?
FILE NAME: 27-2
LIMITED MASS RANGE?: Y
MASS RANGE: 70-350
PRINT?:
EXPAND BY:
8
*
8,
8.
R.
GC-tf3 CCHiD OF SRMFUE 26S32-13-28
Tentative Assignment
A 2,2-dimethyl-3-acetylcyclobutene
carboxylic acid, methyl ester
B cis-pinonic acid, methyl ester
C trans-pinonic acid, methyl ester
ICO 110 120
FIGURE 25. RECONSTRUCTED GAS CHROMATOGRAM OF
a-PINENE REACTION PRODUCTS
(Second Analysis)
-------
SYSTEM 150 IS ON
SELECT MODE: OUTP
PRINT?:
PLOT?.: Y
FILE NAME: 27-2
SPECTRUM NUMBER: 95
AMPLITUDE EXPANSION?:
MINIMUM VALUE %:
SUBTRACT BACKGROUND?: Y
BACKGROUND FILE NAME: 27-2
SPECTRUM NUMBER: 91
BACKGROUND AMPLIFICATION:
NORMALIZE ON:
52
SPECTHJH
3
60-tfS COHO OF
26S32-5S-28
8_
F-
10
'8_
I
.B.
5L
5
«_
8_
D
-«_
O
4
...111. L.I.I. ...l.ll
t, i,
\\ . . ..
w se es 70 eo so
,
i
See Table 3 for explanation of
fragmentation pattern
, J.I 1. ,!
U1
CO
t-i
0
V
*B
t
^N
§
>-*
1 ! ., !
jhHimi] is.' iHjf»^i,.Mj»M|..!.>.in|iii:}.; .,,,), mpn.,^,,,,,,^^.. u.,,,,,,,,,
iea 110 120 139 HO i£Q leg ITS isa isa
FIGURE 26. CHEMICAL IONIZATION MASS SPECTRUM
OF GAS CHROMATOGRAPHIC PEAK (A)
(Refer to Figure 23)
-------
53
Protonated parent-ion of
-methyl "nor-pinonate",
(MH)"1", m/e" - 185.
Fragmentation of Parent-Ion
Process
Fragment
m/e
Loss of H20(MH -18)
167
Loss of CH3OH(MH -32)
153
Loss of HCOOCH (MH -60)
125'
Transannular cleavage
Transannular cleavage
in
99
Transannular cleavage
71
Transannular cleavage
115
TABLE 3 . FRAGMENTATION OF TENTATIVELY IDENTIFIED
METHYL "NOR-PINONATE" UNDER CONDITIONS
OF CHEMICAL IONIZATION (METHANE)
MASS SPECTROMETRY
-------
54
Aerosol in Natural Haze
In light of the above findings, the presence of pinonic acid was
sought in air particulate collected at Smoky Mountains sampling site. The
organic acid fraction of the particulate matter was isolated as described
above. A portion of the isolated acid was treated with diazomethane and the
methyl esters produced were analyzed by GC/MS. The reconstructed gas chromat-
ogram is shown in Figure 27. The mass spectrum shown in Figure 28 identifies
the indicated peak as corresponding to pinonic acid.
The presence of pinonic acid was similarly sought in urban air par-
ticulate collected at the Cooper Union sampling site. No detectible concen-
tration of the compound could be found in air particulate from this source.
In summary, or-pinene, a natural terpene present in forest atmosphere,
was reacted under simulated atmospheric conditions in the smog chamber. Two
reaction products of a-pinene were isolated and identified as pinonic acid, and
"nor-pinoic" acid (2,2-dimethyl-3-acetyl-cyclobutanecarboxylic acid.) Pinonic
acid was found to be present in rural air particulate, but could not be found
in urban air particulate.
Aerosol Generated from 1-Heptene
The aerosol generated from 1-heptene was collected, and the organic
acid fraction was isolated, esterified with diazomethane, and analyzed by GC-MS.
The reconstructed gas chromatogram is shown in Figure 29. Mass spectra of the
compounds corresponding to chromatographic peaks (C) and (D) are shown in Figures
30 and 31. Both spectra exhibit protonated parent ions consistent with methyl
hexenoate, C H COOCH (isomeric, unsaturated acid-esters). Peaks consistent with
loss of methanol and loss of methyl formate are also observed in both of these
spectra. Tentative assignment of these compounds as isomers of methyl hexenoate
*
is reasonable on the basis of the mass spectra.
Mass spectra for chromatographic peaks (A) and (B) exhibit protonated
parent-ions peaks (see Figures 32 and 33) consistent with methyl pentenoate,
C.H COOCH (isomeric, unsaturated acid-esters). Peaks for the parent-ion plus
C_H are also observed. Although both spectra show peaks (very weak) consistent
with loss of methyl formate, neither spectrum shows a peak for loss of methanol.
BATTELUE COLUMBUS
-------
S1TSTKM 150. IS ON
SELECT MODE: IfSS
CALIPRATK?:
TITLE:. PC-MS 01-
CALIBRATION FILE NAME:
FILE NAME: 23-2
MASS RANPE: v0-350
SAMPLES/AMU: 1
MAX RPT COUNT: 4
BASF. INTEGRATION TH-iF: 8
HPT COUNT PEJ-OHE CHECKING
LOWER THRESHOLD: 3
UPPER THRESHOLD: 4
1015 RANGE SETTIVP?:' H
MAX RUU TIME: /)«
55
NATURAL
CAL-H
HA7.K OV-1V.
LOV:FR THP.ESHOLD:
6C-MS
OF NRTURft, HflZE OV-17
0
10 20 30
SPECTRUM NUMBER
FIGURE 27. RECONSTRUCTED GAS CHROMATOGRAMS OF METHYL ESTERS
OF ORGANIC ACIDS ISOLATED FROM NATURAL-HAZE AEROSOL
-------
SYSTEM 15fl IS ON
SELECT tfOOE: OUT? 56
PHI NT?:
PLOT?:. Y
FILE NAME: 23-2
SPKCTP.UK NUKPSK: 34
A.XPH,YUDE EXPAMSI ON?:
MINIMUM VALUE ":
SUETiiACT BACKGROUND?: Y
BACKGROUND FILE NAXF.: 23-2
SPECTRUM \Uli-J52EH: 31
BACKGROUND AK-PLI FI CATI.ON:
NOHXALIZE ON:
SPECTRLM NUMBER 3i
g
8.
GC-MS
OF NflTURFL HfiZE OV-17
-
O
(See Table 1 for explanation
of fragmentation pattern)
Ok
o<
i
0)
"e1
TJTTlTTTTTTpTTTJTn | I I 11 11 I 111 I I I I 111 I 11 I I I I | I I M | 11 I 11 I I 111 I M I 111 I 11 11 11 111 I 11 M I I 11 I 1111 I I I 111 I 11 I | ., . J . . . . | . . . J
30 100 110 120 130 110 ISO 160 170 180 130
60 70 80
M/ E
FIGURE 28
CHEMICAL IONIZATION MASS SPECTRUM OF GAS
CHROMATOGRAPHIC PEAK CORRESPONDING TO
PINONIC ACID
(Refer to Figure 27)
-------
SYSTEM 150 IS ON
SELECT MODS! IFSS
CALIBRATE?!
TITLES GC-MS CCH4) OF HEPTENE - N02 PRODUCTS
CALIBRATION FILE NAME: CAL-M
FILE NAMEl 24-4
MASS RANGES 50-250
SAMPLES/AMUS 1
MAX RPT COUNTS 4
BASE INTEGRATION TIMES 8
RPT COUNT BEFORE CHECKING LOWER THRESHOLDS 4
LOWER THRESHOLDS 3
UPPER THRESHOLDS .4
1015 RANGE SETTING?} M
MAX RUN TIMES 30
GC-HS CCW) OF
- N02 PRODUCTS
l""'""l""
0 10 20 30
SPECTRUM NUMBER
130 Itt
FIGURE 29. RECONSTRUCTED GAS CHROMATOGRAM OF 1-HEPTENE REACTION PRODUCTS
-------
SYSTEM 150 IS ON 58
SELECT MODEl OUTP
PRINT?I
PLOT?: Y
FILE NAME* 24-4
SPECTRUM NUMBER! 52
AMPLITUDE EXPANSION?:
MINIMUM VALUE S»
SUBTRACT BACKGROUND?* Y
BACKGROUND FILE NAME! 24-4
SPECTRUM NUMBER: 48
BACKGROUND AMPLIFICATION:
NORMALIZE ON:
Tentative assignment:
COOCH
methyl hexenoate
SPECTRUM NIHBQR S2
B
GC-MS CCWD OF
- N02 PRODUCTS
o
ON
VO
0
O
o
o
BC
, ,r,lrr,r].,r,nm|rTT,I,Tnrr,r..,,..r1,TTTTpTTT1
33 ICO 110 120 130 110 ISO 160 170 18C
|...,,....|....,...,|....,...,,
10 SO 60 70 I
E
FIGURE 30. CHEMICAL IONIZATION MASS SPECTRUM OF
GAS CHROMATOGRAPHIC PEAK (C)
(Refer to Figure 29 )
-------
SYSTEM 150 IS ON
SELECT MODEI OUTP
PRINT?*
PLOT?* Y
FILE NAME* 84-4
SPECTRUM NUMBER* 63
AMPLITUDE EXPANSION?*
MINIMUM VALUE %*
SUBTRACT BACKGROUND?*
BACKGROUND FILE NAME*
SPECTRUM NUMBER* 59
BACKGROUND AMPLIFICATION*
NORMALIZE ON*
59
Tentative assignment:
methyl hexenoate
Y
24-4
8
8J
SPECTRUM NUMBER 63
OF
0)
~B
- N02 PRCDUaS
bEL
Ox
VO
i
O
8
0
o»
a
PB
0
PI
5
1
SO 60 70 80 39 1(33 110 120 130 1-W ISO 160 170 16C
FIGURE 31. CHEMICAL IONIZATION MASS SPECTRUM OF
GAS CHROMATOGRAPHIC PEAK (D)
(Refer to Figure 29)
-------
OUTP 6°
PRINT?!
PLOT?* Y
FILE NAMEI 24-4
SPECTRUM NUMBER: 2?
AMPLITUDE "EXPANSION?*
MINIMUM VALUE Zt
SUBTRACT BACKGROUND?! Y
BACKGROUND FILE NAME) 24-4
SPECTRUM NUMBER: 21
BACKGROUND AMPLIFICATION!
NORMALIZE ONI
Tentative assignment:
C H COOCH
methyl pentenoate
SPECTRLM NU-BER 27
GO-MS CO-H) OF
- N02 PRODUCTS
in
m
4>
~B
o
g
II
I
0)
CM CO
U O
Ji
SO 60 70 80 S3 103 118 123 133 110 1S3 163 170
IV E
FIGURE 32. CHEMICAL IONIZATION MASS SPECTRUM OF
GAS CHROMATOGRAPHIC PEAK (A)
(Refer to Figure 29)
-------
SYSTEM 150 IS ON 61
SELECT MODE1 OUTP
PRINT?*
PLOT?I Y.
FILE NAME: 24-4
SPECTRUM NUMBER: 39
AMPLITUDE EXPANSION?!
MINIMUM VALUE 5!»
SUBTRACT BACKGROUND?: Y
BACKGROUND FILE NAME: 24-4
SPECTRUM NUMBER: 34
BACKGROUND AMPLIFICATION*
NORMALIZE ONI
Tentative assignment:
methyl pentenoate
SPECTRUM NUMBER 33
GC-HS CCWD OF
- N02 PRODUCTS
10
U
0)
~e
33
1
1
0)
ir\
1-1
n
33
CM
0
4m
X
CO
60 70 80 80 100 110 120 130 110 ISO 160 170
FIGURE 33. CHEMICAL IONIZATION MASS SPECTRUM OF
GAS CHROMATOGRAPHIC PEAK (B)
(Refer to Figure 29)
-------
62
Such a peak is to be expected in the spectra of methyl esters. Thus, the
identification of these two compounds as isomers of methyl pentenoate is
especially tentative.
Various positional isomers of hexenoic or pentenoic acid are pos-
sible. The o>3 unsaturated isomer will display an ultraviolet absorption
band at -^215 nm. The ultraviolet spectrum of the total acid fraction shows
absorption in this region. Thus, the acid mixture probably includes o>0
unsaturated acid.
In summary then, the data tends to indicate that irradiation of 1-
heptene and NO in the smog chamber yields products including mixed hexenoic
X
acids and (possibly) mixed pentenoic acids.
, ,_ (tentative)^
C.H,COOH
4 7
mixed
pentenoic
acids
+ C H COOH
mixed
hexenoic
acids
Aerosol Generated from Toluene
Analysis of auto exhaust particulate revealed the presence of benzoic
acid and phenylacetic acid. In view of this finding, toluene was reacted in the
smog chamber and the acid fraction of the aerosol produced was analyzed. Analysis
failed to reveal a detectable concentration of benzoic acid.
Cross-Contamination in Generation of
Smog Chamber Aerosol
Battelle's 610-cu. ft. environmental chamber was used to generate
aerosol from a variety of pure compounds. In the operation of such a facility,
aerosol is cleared from the chamber at the conclusion of the experiment by
purging with purified air. In this program, purging was carried out for a min-
imum of 12 hours. Our results indicate that such purging may not be sufficient
to remove all aerosol components from the chamber in preparation for subsequent
experiments.
BATTELLE COLUMBUS
-------
63
Aerosol was generated from the several pure compounds listed below
in the following chronological order:
(1) a-pinene
(2) isoprene
(3) cyclohexene
(4) 1-heptene
(5) toluene
As discussed previously in this report, GC-MS analysis of aerosol generated
from a-pinene revealed the presence of pinonic acid. However, the analysis
of aerosols subsequently generated from other precursors similarly revealed
the presence of this acid. Specifically, pinonic acid was found in aerosol
generated from isoprene, cyclohexene, and toluene. Presumably, this crys-
talline nonvolatile acid was not completely removed from the smog chamber
by the purging procedure described above, and thus appears in aerosol sub-
sequently generated from other precursors.
Based on the above data, it appears that the generally accepted
purge procedure employed in this program may not effectively remove some
components of generated aerosol. Alternate procedures should be examined
in future work.
Summary; Organic Reactions in the Smog Chamber
In the analysis of the smog chamber reaction products, the involve-
ment of two general types of reactions have been observed: oxidative cleavage
of an olefinic bond and oxidative decarboxylation. The detailed reaction mech-
anisms involved cannot be postulated at this point. However, in cleavage of
the olefinic bond, formation of a classical ozonide is probably involved rather
than some unspecified free radical mechanism. The oxidative decarbonylation,
on the other hand, may well involve a free radical mechanism.
BATTELLE COLUMBUS
-------
64
Analysis for Polynuclear
Aromatic Hydrocarbons
High-pressure liquid chromatography (HPLC) was evaluated for use
in the separation and analysis of polynuclear aromatic hydrocarbons present
in air and auto exhaust particulate. It was the objective of this phase of
the study to develop a method whereby a rapid quantitative analysis of a
series of PNA's might be conducted. From such data a ratio of PNA's pre-
sent could be calculated. It has been suggested that such a PNA ratio
might be useful in determining the automotive contribution to atmospheric
pollution. However, this effort was discontinued at the direction of the
Task Force in order that more general organic analyses might be carried
out. The results obtained, however, demonstrate the capability of this
technique to separate rapidly a variety of PNA's.
Work was conducted using a Waters Associates Model 100 ALC Auto-
matic Liquid Chromatograph equipped with an ultraviolet photometric detec-
tor. The following materials were used to prepare chromatographic columns
for evaluation.
(1) Oxypropionitrile (OPN) on Porosil C (Waters Associates)
(2) Zipax chromatographic support coated with a hydrocarbon
polymer stationary phase (Du Pont)
(3) Corasil porous glass beads (Waters Associates)
(4) 20 percent Acetylated Cellulose (Machery, Nagel & Company)
(5) CPG 10-240 Corning Glass Beads treated with octadecyl-
trichlorosilane to yield a bonded hydrocarbon stationary
phase.'''
The chromatogram in Figure 34 shows the separation of several
authentic PNA's. The separation was performed using a 1/8-in, stainless
(7) Ledford, C.J., Morie, G. P., Glover, C. A0, "Separation of Polynuclear
Aromatic Hydrocarbons in Cigarette Smoke by High Resolution Liquid
Chromatography", Paper presented at the 23rd Tobacco Chemists' Research
Conference, Philadelphia, Penna. (1969).
BATTELLE COLUMBUS
-------
Authentic Mixture of Aromotic Hydrocorbons
(A) Naphthalene, 0.60 microgrom
(B) Phenanthrene, 0.10 microgram
(C) Pyrene, 0.90 microgram
(D) Benzo(a)pyrene, 0.46 microgram
Column: 9 foot x 1/8 inch, OPN on Porasil C
Mobile Phase-. 0.25 percent methyl-isobutyl ketone
in iso-octane
Flow; 1.3 ml/min
Pressure: 670 psig ' ^x
Chart: 3 min/inch
FIGURE 34. HIGH PRESSURE LIQUID CHROMATOGRAM OF AUTHENTIC POLYNUCLEAR AROMATIC HYDROCARBONS
-------
66
steel column of OPN (oxypropionitrile)on Porasil C with a mobile phase
of 0.25 percent methylisobutyl ketone in iso-octane. The column flow rate
was 1.3 ml/min and the head pressure was 670 psig. Note that the time re-
quired for the separation shown is under 30 minutes and that the sensitivity
obtained is well into the sub-microgram range.
The polynuclear aromatic hydrocarbon fraction of air particulate
collected at the Cooper Union sampling site was analyzed using the above
OPN column. The PNA fraction analyzed was isolated by preparative TLC on
silica gel with chloroform as the developing solvent. The chromatogram in
Figure 35 shows the separation of various components of the PNA fraction.
Based on this analysis, a tentative identification of naphthalene, phenant.h-
rene, and benzo-a-pyrene can be made. Note that the presence of pyrene is
not indicated.
Further evaluation of the above OPN column was conducted. Spec-
ifically, a separation of benzo-a-pyrene and benzo-e-pyrene was attempted.
Satisfactory resolution of the isomeric benzpyrenes was not realized. Thus,
a column was sought which would yield such a separation. Chromatographic
columns prepared from packing materials (2), (3), and (4) above did not re-
solve the benzpyrene isomers; these columns yielded separation of other
PNA's, but were generally inferior to those already obtained using an OPN
column.
A 9 ft x 1/8-inch column was prepared using CPG-10/240 Corning
glass bead treated with octadecyltrichlorosilane (packing material No. 5
above). This column yielded good resolution of benzo-a-pyrene and benzo-
e-pyrene from each other as well as from the lighter PNA's, The column
was used with a mobile phase of acetonitrile at a flow Vate of 0.57 ml/min
and head pressure of 285 psig; the chromatogram is shown in Figure 36.
While giving good resolution of the isomeric benzypyrenes, this
column is not as efficient as OPN on Porasil C for the resolution of the
BATTELLE COLUMBUS
-------
TLC Fraction of Methylene Chloride Extract of Air
. Porticulote Matter, Cooper Union Collection Site
Column: 9 foot x 1/8 inch, OPN on Porasil C
Mobile Phase; 0.25 percent methyl-isobutyl ketone
in iso-octane
Flow: 1.3 ml/min
Pressure: 670 psig
x Chart: 3 min/inch
FIGURE 35. HIGH PRESSURE LIQUID CHROMATOGRAM OF TLC FRACTIONATED METHYLENE CHLORIDE EXTRACT OF AIR PARTICULATE MATTER
-------
I I I I I
Authentic mixture of aromatic hydrocarbons
(A) Aromatic impurities; naphthalene, phenanthrene
(B) Benzo(e)pyrene, 0.9 micrograms
(C) Benzo(a)pyrene, 0.9 micrograms
Column: 3m x 1/8 inch Corning Glass Bead (CPG 10-240)
treated with Octadecyltrichlorosilane
Flow: 0.57 ml/min
Pressure: 285 psig
Detector: UV photometer (540 nm)
Chart speed; 3min/inch
12 IS 18 21 24 27
30
33
36
39
42
; 'FIGURE 36. HIGH PRESSURE LIQUID CHROMATOGRAM OF AUTNENTIC POLYNUCLEAR AROMATIC HYDROCARBONS
-------
69
lighter PNA's. Thus separate analyses using each column should be
performed to obtain maximum analytical data.
INORGANIC ANALYTICAL STUDIES
Bronx Site
One sample, collected on a Millipore filter during a 23-hour
period (9/30/70 to 10/1/70), was analyzed by spark-source mass spectro-
scopy for metals by atomic absorption for lead, and by wet-chemical
methods for N0_ , SO, , and NH, . Results of analyses by spark-source
mass spectrometry and atomic absorption are shown in Table 4; results
of wet-chemical analyses are shown in Table 5. The percentage of
lead is within the expected range. The proportion of sulfate relative to
nitrate is higher than found in Los Angeles reflecting the higher SO.
emissions in New York City. The ammonium concentration is also propor-
tionately higher than found in Los Angeles. On a relative basis the
amount was almost equivalent to the stoichiometric quantity required
for the formation of NH.NO- and (NH.) SO,.
Smoky Mountains Site
Quantitative analysis of aerosols collected in the Smoky
Mountains has not yet been completed. However, sulfur and vanadium
were found but lead was not present.
ATTELLE COLUMBUS
-------
70
TABLE 4. METALS IN BRONX SAMPLES^
Metal
Al
Ca
Si
Fe
Mn
Ba
B
"Mg
Pb
Cr
Pb (by
Percent of Total
Sample Weight
5
5
5
2.5
1
1
0.015
1.5
1.5
0.05
aa) 1.8
Metal
Sn
Ni
Mo
V
Cu
Ag
Zn
Na
Ti
Zr
Percent of Total
Sample Weight
0.15
0.15
0.005
0.5
0.4
0.05
1
1.5
0.35
0.005
(a) All analyses by spark-source mass spectroscopy
accuracy * 50 percent, with exception of Pb which
was also analyzed by atomic absorption (aa).
TABLE 5. WET-CHEMICAL ANALYSIS OF BRONX SAMPLE
Percent of Total Relative Number
Component Sample Weight of Radicals
^4 7.6 1
N03= 13 0.5
S0= 13 0.3
4
BATTELUE COLUMBUS
-------
71
ACKNOWLEDGMENT
The study of the composition of aerosols formed from pure
hydrocarbons was supported in part by Grant AP 00828 from the Office
of Research Grants, Environmental Protection Agency.
BATTELL.E COLUMBUS
------- |