Ecological Research Series
PARTICLE SIZE DISTRIBUTION OF NITRATE
AEROSOLS IN THE LOS ANGELES AIR BASIN
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5. Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems are assessed for their long- and short-term influ-
ences. Investigations include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects. This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-77-053
May 1977
PARTICLE SIZE DISTRIBUTION OF NITRATE AEROSOLS
IN THE LOS ANGELES AIR BASIN
by
Alan Henry Moskowitz
California Institute of Technology
Pasadena, California
Grant No. R802160
Project Officer
William E. Wilson
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
11
-------
ABSTRACT
The atmospheric aerosol was sampled with a low pressure impactor at a
coastal, an urban, and an agricultural site in the Los Angeles air basin.
The material collected on each stage was analyzed for nitrate by direct
vaporization into a chemiluminescent analyzer, sensitive at nanogram levels.
The method responds to inorganic nitrate compounds which vaporize or decompose
below about 1200°C. The coastal nitrate size distribution consists mainly
of particles which have diameters greater than 2.0 ym, whereas the nitrate
in the agricultural region is found primarily in the submicron range. The
urban location, exhibiting characteristics of both coastal and agricultural
regions, was bimodal about the 1-2 ym range. It is believed that the sub-
micron aerosol is ammonium nitrate while the larger size fraction is sodium
nitrate.
This report was submitted in fulfillment of Grant No. R802160-03-6
by California Institute of Technology under the sponsorship of the U.S.
Environmental Protection Agency. This report covers the period October 1,
1975 to December 1, 1976, and work was completed as of December 1, 1976.
iii
-------
CONTENTS
Abstract iii
Figures ..... vi
Tables vi
Acknowledgment vii
1. Introduction 1
2. Apparatus 2
3. Analysis and Calibration . 4
4. Ambient Nitrate Measurements 9
5. Conclusions 16
References . 17
Appendices
A. Nitrate and NO Chemistry 18
B. Techniques and Apparatus 22
C. Ambient Nitrate Measurements 31
-------
FIGURES
Number Page
1 Recovery of nitrate for different compounds. The numbers
to the right of a data point indicate the number of trials
the point represents. For (NH,)2SO,, the nitrate equivalent
to the ammonia content is indicated on the abscissa 5
2 Aerosol nitrate distributions with respect to particle size
for different locations at various times of day 10
3 The average diurnal distribution for aerosol nitrate with
respect to particle size for different locations 11
4 a) Total nitrate as a function of time 13
b) The diurnal variation of the sub-2.0 ym nitrate fraction 13
5 Typical diurnal profile of Los Angeles smog 19
6 Aerosol vaporization apparatus and associated circuitry ... 23
TABLES
1 Comparison of Impactor and Filter Collection of Nitrate ... 8
2 Nitric Oxide Concentrations and Distribution Factors
for Pasadena 14
3 Background and Percent Recovery for Different Coatings ... 29
4 Comparison of Impacted Nitrate using Coated and Uncoated
Strips (All Numbers yg/m3) 30
5 Comparison of Impacted Nitrate using Vaseline Coated and
Uncoated Strips (All Numbers yg/m3) 32
vi
-------
ACKNOWLEDGMENT
I wish to thank my advisor, Professor Sheldon K. Friedlander for his
patience and guidance during my stay at Cal Tech. I also wish to thank
Professor Richard C. Flagan for his help in the earlier phases of this work.
Special thanks is given to all the members of our research group whose
suggestions and advice made this undertaking possible.
vii
-------
SECTION 1
INTRODUCTION
Previous investigators (1-6) have shown that participate nitrate
Compounds, derived from nitric oxide emissions, make up a significant
fraction of the total atmospheric loading in the Los Angeles air basin.
In order to obtain accurate size distribution data for nitrates it is
necessary to use a technique which is capable of measuring small quantities
of nitrate. Wet chemical techniques including extraction, colorimetric
methods, and use of the nitrate specific electrode, often require large
sample sizes and excessive sample handling which increase the risk of
contamination and restrict the lower limit of detectability. An aerosol
vaporization technique is described here which measures nitrate at
nanogram levels directly from the collection substrate. The vaporization
procedure and apparatus are similar to the method described by Roberts
and Friedlander (7) for sulfate. A brief description follows as the
details may be found in that publication.
-------
SECTION 2
APPARATUS
Stainless steel strips one inch long upon which aerosol has been
deposited, are mounted across two tungsten posts in a small glass cell.
A nitrogen flow of 1.2 liters per minute passes into the cell where it
picks up vaporized products resulting from a rapid heating of the steel
strip. The strip, acting as a resistance, is heated by the discharge
of a capacitor (190,000 MFD at 11.6 volts). The vaporized sample is
then carried into a nitrogen oxide-detector.
The detector used is a chemiluminescent N0/N02 analyzer
(Thermoelectron Corp. Model 14B) which is capable of measuring NO or N02
concentrations up to 10 ppm. The NO determination is accomplished by
a gas phase titration of the NO with ozone. The intensity of the
light emitted by the chemiluminescent NO-O^ reaction has been shown to •
vary linearly with NO concentration over many orders of magnitude (8,9).
The instrument is equipped with a molybdenum catalyst operated at 450°C
which, by reducing N02 to NO, allows measurement of N02< Other con-
verters (10) are used to convert organic nitrates such as nitrosoamines
to NO so that these may also be measured. The effectiveness and
sensitivity of the chemiluminescent technique has been demonstrated by
its use in conjunction with a gas chromatograph. N-nitroso compounds
at the sub-ppb level have been quantitatively identified in air, water
and foodstuffs (11).
The chemiluminescent analyzer is calibrated by using an NO
calibration gas produced by a dilution system (Kin-Tek Precision
Calibration System Model 570). A flow of nitric oxide in a nitrogen
diluent is produced in this system by a permeation membrane which
produces calibration gases having concentrations of a few tenths of a
ppm.
-------
The vaporization of nitrate results in the formation of NO; less
than r/o of the nitrate is converted to N02. However, in order to be
certain that all of the nitrogen oxides are measured the catalyst-is
used and all oxides of nitrogen are monitored. The vaporization may be
carried out in air; tests with zero air yield the same results. The
chemiluminescent analyzer responds rapidly to the NO pulse sent from the
vaporization apparatus. The response time of the analyzer is short
enough to permit quantitative NO detection. A scale setting on the
analyzer must be chosen so that the amplitude of the output signal
(ppm NO ) will lie between 5 and 95 percent of full scale in order
A
to assure accuracy and avoid saturation of the instrument. Generally
a full scale setting of 2 ppm is employed which provides a maximum
input pulse of 250 ng of nitrate. At this setting all output signals
have accurately measurable traces. The quantity of nitrate vaporized
is related to the output trace by means of the equation:
Mass N03 = (Trace area)(MW of N03)(Flow in 1/min)(Atmospheric P) (1)
(K in counts/ppm-min)(R)(T in °K)(106ppm)
where the area is given in counts; K, a constant determined from the
scale chosen for the integrator; R, the gas constant in liter-atm/mole-°K;
the atmospheric pressure in atm. An electronic integrator (Varian
Aerograph Model 477 Electronic Integrator) determines the area under
each pulse . The input flow rate is held constant by the detector.
-------
SECTION 3
ANALYSIS AND CALIBRATION
Stainless steel strips were baked in a furnace for 1 hour at 900 C,
A sample of 23 strips was examined for background levels. It was found
that the strips contained 1.7 ng of N (standard deviation 0.63 ng),
equivalently 5.6 ng N02 or 7.5 ng NOg. Furthermore, it was discovered
that after the strips were blanked by capacitor discharge, adsorption
occurred upon exposure to ambient air. About 0.5 ng of NOV, calculated
' ' A
as N, is adsorbed in 1 minute. It was determined that a limit of
approximately 2 ng of NO , as N, is reached in a few hours. This falls
A
within the range of the average strip background and suggests that the
adsorbed species is responsible for the background. No correlation was
found between.the ambient NO levels and the quantity, of gas adsorbed.
A
It is believed that the history of the strip surface is the more
important factor in explaining background levels. Because of adsorption
the lower limit of quantitative nitrate analysis is about 25 ng of NO.,.
A non-adsorbing surface would increase the sensitivity of the method.
Standard solutions are placed on preflashed strips using a 1
microliter syringe, and are allowed to dry in the nitrogen stream.
Solutions of NH4N03, NaN03, Pb(N03)2, (NH4)2S04 and HN03 were prepared
using doubly distilled water and predried salts. Figure 1 shows the
measured NO, calculated as nitrate, vs. the quantity of nitrate placed
on the strip. By considering the ammonium nitrogen to have a nitrate
equivalent on the abscissa, (NH.^SO^ is included to show that
ammonium ion is not detected. The low levels observed for (NH.)2S04
are due to impurities in the standards. The standard deviation
(excluding HN03 and (NH^SO.) is 15%, predominantly arising from the
preparation of the standards rather than from the technique itself.
The HN03 recovery is poor since its low boiling point and high
-------
480
UJ
Q
X
o
o
cr
400
UJ
< 320
o:
h-
"2 240
Q
LJ
CO
<
LJ
Q
LJ
O
_l
<
O
160
80
0
NH4N03 •
NaN03 B
Pb(N03)2 *
(NH4)2S04 o
HN03 a
B
8
B'2'
D
8
0 80 160 240 320 400 480
QUANTITY OF NITRATE PLACED ON STRIP, ng
Figure 1. Recovery of nitrate for different compounds. The numbers
to right of a data point indicate the number of trials it
represents. For (NH/^gSO^, the nitrate equivalent to the
ammonia content is indicated on the abscissa.
-------
volatility promote its evaporation from the strip. When using large
aliquots of HNOg the detector actually responds to a pulse of NOX prior
to the heating of the strip. All of the salts investigated have
vaporization or decomposition points under 500 C, far below the 1200 C
temperatures to which the strips are exposed. Our studies show that
under or over baking the strips by as little as half an hour sharply
reduces the recovery of nitrate after vaporization.
The stainless steel strips are mounted in a single jet, 8 stage,
low pressure impactor which is capable of size fractionating aerosols in
the submicron range; lower diameter cutoffs, in microns, are 4, 2, 1,
0.5, 0.25, 0.11, 0.05, 0.02 (12) Lundgren (2) reported that the
aerosol collected on the stages of an impactor may be 20-30% less than
that collected by a total filter. This is especially true when solid
particulate matter as opposed to watery aerosol is collected. An attempt
was made to determine exactly how much nitrate was passing through the
impactor by placing a glass fiber filter after the last stage of the
impactor. Since good quantitative nitrate recovery was possible by
sandwiching a piece of the filter (0.25 cm in diameter) between 2 strips
and heating at 25 volts, it seemed possible to measure any nitrates not
collected in the impactor. This approach was abandoned as the nitrate
levels on the after filter could not be measured with sufficient
accuracy to permit quantitative analysis. The levels were almost the
same as that of the filter background due to the low volume of gas
passing through the filter. In addition, the standard deviation for
a given filter was about 25% of the total nitrate present and tended
to obscure any results. It was then assumed that bounce off occurred and
we chose to find a suitable coating for the strips.
The best coating would be a compound which is sticky, nonvolatile,
nitrate free, and does not combine with the highly reactive NO that is
evolved during the sample heating. Various substances were tried in-
cluding glycerol, silicone grease, mineral oil and vaseline. Vaseline
was the most suitable although the blank levels were not always
negligible. The vaseline is applied with a cotton swab in & very thin
layer. The nitrate level on the coated strip is 12 ng NO.,, standard
-------
deviation Is 5 ng. Studies done with ambient aerosol indicate that
there is a 45-65 percent increase in the total nitrate collected when
using the coated strips as compared to the uncoated ones. It was
possible to compare total nitrate collected by the impactor with that
collected by a glass fiber filter (Gelman Type AE). In all cases the
level of nitrate on the after filter was not measureable, due to
the aforementioned difficulties. Table 1 shows the data for 3
comparisons. The flow through the impactor is 1 liter per minute; that
through the filter is 44 liters per minute. Within experimental error
the nitrate levels obtained by both methods were identical.
-------
TABLE 1. COMPARISON OF IMPACTOR AND FILTER
COLLECTION OF NITRATE
Nitrate collected by Nitrate collected by Duration of
impactor, ^.g/m3 filter, /xg/m3* sample, minutes
14.0 13.6 60
6.42 7.3 80
47.8 41.1 25
^Standard deviation is 25%
-------
SECTION 4
AMBIENT NITRATE MEASUREMENTS
Measurements of aerosol nitrate were made at three locations in the
South Coast air basin; Pasadena, Hermosa Beach (30 miles SW of Pasadena),
and Chino (an agricultural area 30 miles SE of Pasadena). The 8 stage,
low pressure impactor sampled at a flow rate of 1 liter per minute.
Six samples were obtained over the period 7 AM to 7 PM with each sample
being 1-2 hours in duration. The strips were analyzed the same day.
The data for the three locations for each run are shown in
Figure 2. It is seen that in midmorning, regardless of location, there
is a significant quantity of particulate matter in the submicron range.
This is also apparent in the late afternoon data of Chino and Pasadena,
suggesting that this size range may well be associated with auto exhaust.
An average diurnal nitrate size distribution for each location
was obtained by time averaging the nitrate levels present on like stages
of the impactor. The distributions are shown in Figure 3. The Pasadena
distribution is bimodal; peaks exist in the two size ranges .05-1 Mm and
2-8 ym ( 8 ym is an arbitrary upper cutoff). The Chino experiment, near
an ammonia rich cattle feed area, possesses a significant peak in the
submicron range although the distribution appears to be weakly bimodal.
In the coastal area the predominant size range for nitrates is in the
2-8 ym region. Pasadena may be thought of as possessing aerosol
characteristics of both coastal and agricultural regions.
It is of interest to compare these results with those of other
investigators. Lundgren (2) reported that on days of heavy smog in
the eastern region of the basin (Riverside), a large amount of
crystalline, hygroscopic particulate matter was present in the .5-1 ym
range. X-ray diffraction identified it as NH.NO.,. Infrared spectra
studies by Grosjean (1) show that NH.NO., comprises 95% of the total
-------
10.0-
HERMOSA BEACH
10/4/76
PASADENA
10/5/76
CHINO
10/8/76
ro
0-
7.5-
5.0-
2.5-
0-
7.5-
5.0-
2.5-
en
_o
^
f.
O
0-
7.5-
5.0-
2.5-
6:42-8:30
I I I
8:43-10:31
I I I I I I I I
10:43-12:30
I I I
12:41-14:47
I I Tl T
15:00-16:30
6:58-8:47
ii I I I I I
9:15-10:29
\ I I I I I I
10:37-12:12
I I I I
12:25-13:45
\ \ II T
15:14-16:10
i i I i i i i i
16:42-18:28
ii i i i i i
17:15-19:01
7:05-8:30
I I I I I I I
8:42-10:12
I I I I I I I
10:37-12:00
1II
12:12-14:00
I I i 1 T
14:12-16:01
i I l l l l l I i I I l i i l iiiiiii
.02 .05 .11 .25.50 I 2 4 8 .02 .05.11 .25.50 I 2 4 8 .02 .05 .11 .25.50 1248
dp,/xm
Figure 2. Aerosol nitrate distributions with respect to particle
size for different locations at various times of day.
10
-------
5
ro
v^ 4
en
Q.
O
en
_O
ro 2
O
<3 ,
1
—
0.(
_=___
D2 0.(
rosQQenu
Chino
r~> _ _ _ L.
Beach
— — J
_.- 1
1
D5 0
_-J
| '
!
.11 o.;
i 1
25 0
;
.5 1.
1
•
4
1 1
BB OH» ^BIB •••»• ••B
0 2.
_- 1
0 4
1
.0 8
.0
Figure 3. The average diurnal distribution for aerosol nitrate with respect to particle
size for different locations.
-------
nitrate aerosol. In an investigation by O'Brien (6), inorganic nitrate
and other participate pollutants were measured in a number of California
locations. It was found that inorganic nitrate (NH»N03 and NaNCL) of a
secondary or photochemical origin contributed about 10% of the mass
loading in the South Coast air basin. The prevalent form of nitrate in
the basin was NH^NO.^ the NH^ concentration was the highest in the
Los Angeles air basin compared to other California locations. The study
also showed that the coastal Santa Barbara sample contained 16 %
inorganic nitrate while the NH. concentration was essentially zero.
This indicates that the reaction of N02 with Nad is of importance in
coastal areas. The mechanism has been postulated to be (13):
3N02 + H20 = 2HN03 + NO (2)
HN03 + NaCl = NaN03 + HC1 (3)
The data indicate that the Pasadena aerosol consists of both
small NhLN03 and larger NaN03 particles. The data from each location
were grouped into 2 categories: particles with diameter greater than
2.0 urn and those with diameter less than 2.0 ym. The percentage of the
total nitrate existing"as small particles for each location was plotted
as a function of time (Figure 4a). The total nitrate as a function of
time is included (Figure 4b). Trie data for Pasadena and Chi no show that
the percentage of nitrate as small particles is reasonably constant over
the day. In the case of the beach it can be noted that the small
particle fraction is significant in the morning, while the larger
marine aerosol dominates later in the day.
Nitrogen gas-particle distribution factors were calculated for
Pasadena experiments. The distribution factor f^ is defined as:
fN = N03" (4)
NOV + NO '
A 0
3
where NO, = particulate nitrate concentration, as N09, yg/m and NO =
•J L. o X
gas phase concentration as N02, converted from ppm to yg/m , averaged
over the sampling period. The data are presented in Table 2. .The
distribution factors are in the same range reported by Grosjean. It
12
-------
LJ
o
UJ
12
10
E 8
V.
5. 6
4
2
E 80
cr CM
t v
eo
U_
•o 40
20
0
© Pasadena 10/5/76
- o Beach 10/4/76
- x Chino(Agricultural Area) 10/8/76
Figure 4b
I
800
1000
1200
1400
1600
TIME
1800
Figure 4a. Total nitrate as a function of time.
Figure 4b. The diurnal variation of the sub-2.0 \im nitrate fraction.
-------
TABLE 2. NITRIC OXIDE CONCENTRATIONS AND
DISTRIBUTION FACTORS FOR PASADENA
NOx(asN02) NOjtas NO 2)
TIME* NOX (ppm) /zg/m3 ^g/m3 fN
Pasadena 9/30/76
9:57 .047 52.0 2.80 .041
11:00 .031 33.2 3.85 .067
12:44 .028 30.0 2.80 .056
14:15 .027 28.9 2.39 .044
15:48 .031 ' 33.2 1.81 .028
17:15 .043 46.0 2.37 .025
Pasadena 10/5/76
7:23 .150 161 4.57 .028
9:52 .075 80.2 7.38 .084
11:24 .031 33.2 5.22 .136
13:05 .020 21.4 . 5.37 .200
15:42 .032 34.2 5.91 .147
18:08 .020 21.4 6.31 .228
*The time shown is the midpoint of the sample interval
-------
would be very interesting to monitor NO levels at the beach and
X
agricultural areas to determine if the abundance of stabilizing cations
(NhL , Na ) actually increases the distribution factor.
It is of interest to look at the amounts of NaNOo which would
correspond to the ambient levels of Na. Sodium was measured in the
ACHEX experiment (14) and the level was determined to be roughly
•3 3
0.7 yg/m . This corresponds to a level of 1.89 Ug/m of N03 . If one
time averages the levels of particles with diameter greater than
2.0 ym for each of the locations it is found that beach nitrate is
333
2.6 yg/m , Pasadena 2.4 yg/m , and Chino 1.0 yg/m . This would be
consistent with the hypothesis that the concentration of Na should
decrease as one gets farther inland, with decreasing levels of large
particles. It is seen that by a rough approximation the concentration
of large particles in Pasadena is consistent with the concentration
predicted by ambient Na levels. One might expect that other anions such
-2
as SO. and Cl would form stable sodium salts also.
Visibility degradation has been correlated with high aerosol
mass concentration. As a result of the ACHEX study it was determined
that nitrate is a less efficient light scatterer than sulfate. Presumably
this is due to the fact that all of the sulfate exists in a highly
efficient light scattering range (.1-1.0 ym) whereas only part of the
nitrate is found in this range.
15
-------
SECTION 5
CONCLUSIONS
A method has been developed which permits the fractionation and
analysis of inorganic nitrate aerosols with a minimum of sample handling.
Although the vaporization and chemiluminescent detection of nitrate is
(accurate, the vaseline coating restricts the lower limit of nitrate
:analysis to 25 ng, which may correspond to a relative error of as much
jas 25%.
Nitrate size distributions vary dramatically with location.
Marine aerosol, present in the greater than 2.0 ym fraction, decreases
i
I in concentration as one gets farther inland. NaNO-. is the principal
'nitrate species in afternoon coastal aerosol. In the inland agricultural
area of Chino, the level of marine aerosol is low. While ammonium
'nitrate aerosol in the sub-2.0 urn fraction is the predominant nitrate
:species in Chino, it exerts only a small perturbation on the coastal
nitrate size distribution. The mixture of aerosols from both coastal
and agricultural regions is exhibited in the bimodal nitrate size
distribution of Pasadena.
16
-------
Chemiluminescent Reaction of NO with O^" Ford Motor Co.
REFERENCES
1. Grosjean.D., Fried!ander.S.K. J. Air Pollution Control Assoc.
25, 1038 (1975)
2. Lundgren, D.A., J. Air Pollution Control Assoc., 20, 603 (1970)
3. Novakov.T., Mueller, P.K., Alcocer,A.E., Otvos.J.W., J. Colloid
and Interface Science, 39,225 (1972)
4. Sandberg.J.S., Levaggi.D.A., DeMandel, R.E., Siu.W., J. Air
Pollution Control Assoc., 26, 559 (1976)
5. Gordon,R.J., Bryan,R.J., Env. Science and Tech., 7, 645 (1973)
6. O'Brien, R.J., Crabtree.J.H., Holmes,J.R., Hoggan,M.C., Bockian,
A.H., Env. Science and Tech. 9, 577 (1975)
7. Roberts,P.T., Friedlander.S.K., Atmospheric Env., 10, 403 (1976)
8. Fontijn.A., Ronco,R.5 Sabadell, A. Anal. Chem., 42, 575 (1970)
9. Stuhl.F., Niki,H.,"An Optical Detection Method for NO by the
Chemiluminescent Reaction of NO
Technical Report SR70-42 (1970)
10. Sigsby.J.E., Black,F.M., Bellar, T.A., Klosterman, D.L., Env.
Science and Tech. 7, 51 (1973)
11. Fine.D.H., Rounbehler.D.P., J. of Chromatography, 109,271 (1975)
12. Hering, S.V., Private Communication
13. Robbins.R.C., Cadle.R.D., Eckhardt.D.L., J. Meteorology, 16, 53
(1959)
14. Hidy, G.M. (ed), Characterization of Aerosols in California
(ACHEX). Final Report to Air Resources Board, State of
California, ARB Contract No. 358 (1975)
15. Calvert, J.G.."Interactions of Air Pollutants" in Proceedings
of the Conference on Health Effects of Air Pollutants. U.S.
Gov't. Printing Office Serial No. 93-15 (1973)
16. Seinfeld, J.H.,"Air Pollution, Physical and Chemical Fundamentals"
McGraw Hill, NY,NY. (1975)
17. Niki.H., Daby.E.E., Weinstock,B., Chapter 2 in "Photochemical
Smog and Ozone Reactions", F.F. Gould (ed), Advances in
Chemistry, Series 113, ACS, Washington, D.C. 1972, p. 16.
18. Wilson,W.E., Ward.G.F., "The Role of CO in Photochemical Smog"
Presented at 160 th meeting of A.C.S., Chicago, 111., Sept. 1970
19. Jacobs, M.B.,"The Chemical Analysis of Air Pollutants",Interscience,
NY,NY (1960). p. 413.
20. Thrush, B.A.,etal Transactions of Faraday Society, 60, 359 (1964)
21. Hodgenson, etal."Applications of Chemiluminescence" in Stevens and
Herget (eds) Analytical Methods Applied to Air Pollution Measure-
ments., Ann Arbor, Michigan (1974)
17
-------
APPENDIX A
Nitrate and NO Chemistry
A
The NO (NO + N09) that usually enters the urban atmosphere is
A . £
primarily NO which is formed during combustion. Although some N02 may
be formed via the overall reaction:
2NO + 02 = 2N02 (1)
most of it is produced by the reaction of NO with ozone, the latter
formed in the atmosphere by the action of sunlight on Op. N02 is also
dissociated by sunlight. The predominant processes for nitrate
chemistry are the following reactions:
N02 + hv = NO + 0 (2)
0 + 02 + M = 03 + M (3)
03 + NO = N02 + 02 (4)
Other reactions produce intermediate species and have been shown to
have some significance (15). These reactions are:
03 + N02 = N03 + 02 (5)
N03 + N02 = N205 (6)
N2°5 + H2° = 2HN03 (7)
However, these reactions alone do not explain the typical diurnal
pattern observed for these species. Whereas one would expect the
concentration of N0? to be diminished as sunlight irradiation
continued, with a steady level of 03 produced rapidly, actually
substantial amounts of 0., are produced and increase dramatically.
Referring to Figure 5, the ozone peak corresponds to the peak in solar
intensity. NO concentration correlates with the morning rush hour
traffic as one would expect.
18
-------
O.lSf—
I
Hour of Day
Figure 5. Typical diurnal profile of Los Angeles smog.
19
-------
In order for the ozone level to peak, it cannot be used in
oxidizing NO to N02> A number of researchers have demonstrated (17, 18)
that other species can contribute to the oxidation of NO without the
consumption of ozone. One such mechanism involves CO. Nitrous acid,
formed in the overall reaction:
NO + N02 + H20 = 2HN02
photolyzes according to:
HN02 + hv = NO + OH (9)
and generates a minor source of the free hydroxyl radical. These
hydroxyl radicals react with CO, also a combustion product, to yield
C02 plus a hydrogen free radical, which reacts with 02 to give the
hydroperoxyl radical, H0« (16):
OH + CO = H + C02 (10)
H + 02 + M = H02 + M (11)
H02 + NO = N02 + OH (12)
OH + CO = H + CO, etc. (10)
o t
Thus, N02 is formed without the consumption of ozone. However, this
mechanism does not entirely account for the levels or the rates
required to explain the patterns observed in the atmosphere. Various
organics introduced into the atmosphere during incomplete combustion
of fuel generate the H02 radical. Olefins are particularly susceptible
to attack by ozone and generate many acylperoxy (RQO?) or alkylperoxv (R0?)
0
compounds which are capable of acting as oxidizing agents. Peroxyacetyl-
nitrate (PAN) is a common, stable compound which is formed by:
CHo-C-0-0 + N09 = CH.-C-O-O-NO- . (13)
j H c. 3 c.
This reaction tends to lower the amount of free NOp available for
generating inorganic nitrates such as NH.NO. and NaNO~.
The relationship between the nitric oxides and particulate
nitrate is not well understood. It is assumed that nitric acid formed
in reaction (7) reacts with NH3 for example, yielding NH.NOo. The NH.
20
-------
may well be solvated prior to the reaction. NH~ is introduced into the
atmosphere by means of a variety of sources including auto exhaust,
industrial exhaust, and evaporation from cattle feed lots. Urban
ammonia concentrations have been estimated to be up to 0.2 ppm with an
average of 0.02 ppm (19). As previously stated, NH.NO, has been
determined to be the major particulate nitrate in the Los Angeles area,
accounting for up to 95% of all particulate nitrates.
The study by Novakovset al. (3) is interesting because pyridino
and amino nitrogen were detected primarily in the small, .6 to 2.0 ym
diameter size range, whereas nitrates were found to be present in only
the 2 to 5 ym range. The amino and pyridino nitrogen are known to be
contained in gasoline additives.
Generally, researchers in the Los Angeles area, including the
eastern part of the basin, report concentrations of nitrate between
3 3
5 and 50 yg/m > with an average of about 7 yg/m . The levels in San
Francisco, for example, are lower, the average being 2.8 yg/m over a
5 year period (4). There is still considerable question as to the size
range of nitrate aerosol, which has been the subject of this paper.
21
-------
APPENDIX B
Techniques and Apparatus
VAPORIZATION APPARATUS
Figure 6 shows a diagram of the glass vaporization cell and the
associated circuitry. It is important to keep the stainless steel
washers free of dirt; they are periodically cleaned with fine sandpaper.
The strips, .8 in. x .21 in., are cut from .001 inch thick stainless
steel shim stock, type 302 full hard. Holes are punched in the strips
so as to fit the tungsten posts. The strips are wiped clean with a
lint free cloth and heated in a furnace for 1 hour at 900°C. Low nitrate
background levels are produced in this way.
*
CHEMILUMINESCENT DETECTOR
The chemiluminescent NO analyzer is only one of the many
J\
methods available used to measure NO and N02 concentrations. It is -
however, the easiest to use and is the most accurate since it does not
require the dissolution of ambient gases into a solvent with the
attendant wet chemical methods. The analysis is performed entirely in
the gas phase. The method takes advantage of the gas phase
chemiluminescent reaction between NO and 0.,; the mechanism has been
shown to be:
kl
NO + 03 =' N02 + 02 (14)
NO + 03 =2 N02*+ 02 (15)
N02* =3 N02 + hv (16)
N02* + M =4 N02 + M (17)
22
-------
To Detector
/ \
U
Nitrogen In
Gloss Cell
Washers
Stainless Steel
Strip
Tungsten Posts
, lr _n X%
Figure 6. Aerosol vaporization apparatus and associated circuitry.
23
-------
where M is any gaseous species present and N02 represents an electroni-
cally excited N02 molecule. The activation energies for reactions 14
and 15, are 2.4 and 4.2 kcal/mole, respectively. Approximately 10% of
the reacting NO follows reaction 15 (20). The reaction between NO
and Oo is bimolecular and the energy of the excited state is about
40 kcal/mole. The intensity can be related to the concentration of the
reactant species as follows:
*
We assume that N02 is in steady state; and making that
approximation we have the following: (18)
d(N02 ) = k2 (N0)(03) - k3 (N02 ) - k4 (N02 )(M) = 0
dt
(N02*) = k2(NO)(Q3) (19)
k4(M) + k3
In designing the system the concentration of M is kept as low as
possible by using a reaction chamber at low pressure; k., becomes the
more significant factor, yielding:
*
Intensity I = K(N02*) = k2(NO)(03)
"3
Fontijn (8) was the first to look into the possible use of
this reaction as a method for measuring the amount of NO present
in the atmosphere. Certainly the predominant reason for his
investigation lies in the specificity of the wavelength of the light
emitted in the reaction. With the proper selection of filters only the
desired reaction will be monitored. Fontijn 's system basically
consisted of a one liter glass reaction chamber into which the ozone
(from an 03 generator) and NO are separately added. A vacuum pump kept
the reaction chamber pressure at one torr. A photomultiplier tube with
an appropriate filter was used to monitor the reaction. One of the
most significant results of this work was to show that the method is
24
-------
highly linear; that is the light intensity varies linearly with NO
concentration over the range 4 ppb to 100 ppm. This represents the range
investigated rather than the limits of reaction linearity. Also of
great importance was the fact that the introduction of gases found in the
polluted atmosphere (at higher than normal concentrations) had no
significant effect on the reaction. Gases tried were NOp-9 ppm, COp-SOO
ppm, CO- 300 ppm, C2H4-5 ppm, NH3-9 ppm, S02-25%, H20-25% saturation.
About a year after Fontijn's paper, Stuhl and Niki (9) at Ford
essentially duplicated Fontijn's work although they made some variations
in the experimental setup. One of their important observations was that
*
at sufficiently high partial pressures of NO and 0,, the N02 intensity
became less susceptible to small pressure fluctuations. They believed
that this was the effect of balancing the rate of formation and
*
deactivation of N02 .
Another possible reaction to use for the measurement of NO is:
NO + 0 + M = N02* + M (22)
N02 = N02 + hv .4ym
-------
increasing temperature. Dissociation begins at 150°C and is 100%
completely dissociated at 620°C. The rate of NO oxidation to N0«
decreases with increasing temperature. In the case that CO and
hydrocarbon levels are much greater than the concentration of oxygen,
,NO can be reduced to H9 in the converter. Converters have been made
A £•
of various metals, including copper and gold, and typically have
efficiencies greater than 95%.
One drawback to the use of the stainless steel converter is that
NHL is also oxidized to NO, provided that oxygen is present.
4NH- + 50-------- > 4NO + 6H?0 (26)
J ^catalyst *
The NhL response has been separated from NO and N0~ by a subtraction
-3 2
technique (10). A tube containing acidic Cr20y~ on firebrick is
placed in the gas stream just ahead of the entrance to the converter.
The following reaction occurs: (27)
2NH3 + Cr20?"2 + 2H+ >(NH4)2Cr207 > N2 + Cr203 + 4H20
Thus only the oxides of nitrogen are measured when the subtractive
tube is used. By the difference measurement, the amount of NhL
present in an ambient sample can be determined. The chemiluminescence
method exhibits many desirable characteristics- selective response to
NO and N02, wide range of use, accuracy and repeatability, minimal
interference by common pollutants, good response time, simple sample
preparation and reasonable low cost.
In the Thermoelectron Chemiluminescent Analyzer Model 14B, the
catalytic converter is molybdenum and is only sinsitive to N02. Thus
organic nitrates such as PAN are not detected. The full scale ranges
on the instrument vary from .05 to 10 ppm and the instrument can
automatically cycle between measuring NO and NO + N02. The upper limit
of detection is about 2000 ng of nitrate in a single vaporization. This
can be increased by diluting the input pulse so that the detector
receives a broader but lower amplitude signal. It is important that
the gas entering the detector be at atmospheric pressure and in order
to connect the flow to a cylinder of nitrogen, an atmospheric dump
bypass is used; in that way the flow rate required by the detector is
26
-------
matched exactly. The analyzer was periodically calibrated. Although
calibration gases of a few tenths of a ppm in nitrogen have recently
been developed by Airco Corporation, the method which we had available
used a gas dilution system, as described below.
KIN-TEK PRECISION GAS CALIBRATION SYSTEM MODEL 570
Calibration of the detector requires a calibration gas consisting
of no more than a few ppm NO in N~ fed directly into the detector. The
Kin-Tek is a system whereby a raw gas, in this case pure NO, is emitted
into a stream of dilution gas (N~). This is accomplished by means of a
permeation membrane.
A chamber containing NO at a nominal pressure is separated from
the nitrogen stream by a semipermeable membrane. The rate of flow of
NO through the membrane as a function of temperature has been measured
by the manufacturer. The temperature of the entire dilution chamber is
held constant by a thermostat and any of three temperature set points
may be chosen. The flow of nitrogen was measured with a wet test meter
and was found to be 1.18 1/min at atmospheric pressure (usually 740 mm Hg),
The usual procedure was to add raw gas and wait until thermal
equilibrium was reached at 30°C. All pressures were then set. The
chemiluminescent analyzer monitored the concentration of the calibration
gas until a constant composition was indicated. At this time the
detector is set to read the proper concentration, then a second
temperature set point is selected (60°C). When equilibrium is established
again the difference between the observed and calculated readings was
never more than .01 ppm. It usually took about 6 hours for the
concentration of the calibration gas to reach a constant value. At 30°C
and 60 C the concentrations of the calibration gases were .32 ppm and
.71 ppm, respectively. The zero point on the detector was set prior to
calibration by flowing zero air into the detector.while monitoring NO.
It was assumed that any NO present in the zero air would be converted
into N02 by the 0^ so that the concentration of NO in the zero air
cylinder could be taken as zero. The cylinders of nitrogen typically
contained .02 ppm NO ; this was taken into account during calibration.
A
27
-------
All procedures were followed as per the Kin-Tek instruction manual and
all connections were made with teflon tubing and stainless steel
:Swagelok fittings.
USE OF STAINLESS STEEL CONVERTER
An attempt was made to use a different converter to obtain
additional information about atmospheric composition. A stainless
steel converter heated to 750°C from a higher range chemiluminescent
detector was connected in series to the 450°C molybdenum converter in
our own instrument. Other investigators (10) have reported that the
stainless steel converter will convert NO,,, NhL, PAN into NO, thus
allowing these species to be detected. The plan was then to use the
subtractive tube to absorb NhL and hopefully not PAN. Then, since the
i ^
molybdenum converter can only detect N02, one could obtain by difference
the concentrations of each of the four species. Unfortunately the
method did not work; two explanations are possible. One is that because
the signal had been detected after a long delay it was believed that
the converter was very long. The longer it was the more it tended to
broaden the pulse and lower the amplitude, resulting in poorer detection.
The other explanation is that since the stainless steel converter was
designed for an instrument capable of measuring 1-10,000 ppm NO , the
^
converter itself may have adsorbed the NO pulse on its interior walls.
COATINGS
We studied the recovery of standards with various coatings as
indicated in Table 3. It must be noted that the percent recovery
with standards is not a conclusive measure of what the recovery will
be from impactor samples. Whereas the impactor samples will consist of
solid particles sticking to the coating, the procedure for testing with
standards involves placing the aqueous standard solutions on the coating
and relying on evaporation to leave the nitrate residue.
The first coating studied in the impactor was glycerol. Looking
at the values in Table 4 it is seen that the coated strips do not
always yield higher nitrate values. The major difficulty in using
glycerol was that because it is a low viscosity liquid the air jet of
28
-------
TABLE 3. BACKGROUND AND PERCENT RECOVERY
FOR DIFFERENT COATINGS
COATING
GLYCERIN
SILICONE GREASE
GE SILICONE OIL
MINERAL OIL
VASELINE
BACKGROUND,
ng N03
14
II
10
15
12
RECOVERY OF
STANDARD, %
62
31
15
86
98
29
-------
TABLE 4. COMPARISON OF IMPACTED NITRATE USING 6LYCEROL COATED AND UNCOATED
STRIPS (ALL NUMBERS yg/m3)
u>
o
8/4/76 11:49 -12:35
Coated Stage Uncoated
39% Increase
8/5/76 10:45-11:45
Coated Stage Uncoated
3.5% Increase
8/5/76 14:08-14:38
Coated Stage Uncoated
5.94
1.67
.43
.40
8.44
3
4
5
6
4.15
1.08
.27
.58
6.08
6.69
1.56
.26
.07
8.58
3
4
5
6
7.04
.79
.12
.34
8.29
5.11
4.16
.26
.34
9.87
1
3
5
6
3.64
6,09
.06
.35
10,14
2.7% Decrease
-------
the impactor forced the coating toward the edges of the strip. This
caused nitrate to be displaced to areas of the strip which are not heated
and at the same time left the central strip area unprotected against
bounce off. The vaseline coating is seen in Table 5 to be consistently
superior to the uncoated strip.
31
-------
TABLE 5. COMPARISON OF IMPACTED NITRATE USING VASELINE COATED AND UNCOATED
STRIPS (ALL NUMBERS yg/m3)
CO
ro
8/10/76 11:18- 12:00
Coated Stage Uncoated
Coated Stage Uncoated
I 1.38
3 2.32
5 .1,2
6 .1,8
4.00
62% Increase
8/11/76 9:28-I P.: 45
Coated Stage Uncoated
8/12/76 11:28-13115
Coated Stage Uncoated
3,71
4.99
2.30
1.07
12.07
40%
8/10/76
1
3
5
6
2
5
8
.33
.14
.69
.46
.62
1
2
1
6
Increase
14:
.48
.24
.85
.47
.04
58%
I
3 1
4
6
3
Increase
.89
.87
.74
.33
.83
13-15:55
3.24
4.08
2.76
2.55
2.78
2,20
.33
0
1
2
3
4
5
6
7.
8
1
2
I
2
1
2
.26
.16
.01
.84
.59
.97
.47
.14
17.94 ,12.44
44% Increase
-------
APPENDIX C
Ambient Nitrate Measurements
All measurements were made during the same week under roughly the
same meteorological conditions. The Hermosa Beach data were obtained
from a balcony, one story above ground level (about 30 feet above sea
level) at a distance of about one-quarter mile from the surf. Wind of
not more than a few MPH was generally from the ocean. Pasadena
meadurements were made from the roof of Keck Laboratory. Measurements
at Chino were made at 3 feet above ground level on the grounds of the
State Correction Facility there, located adjacent to cattle feed lots.
33
-------
TECHNICAL RETORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/3-77-053
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
PARTICLE SIZE DISTRIBUTION OF NITRATE AEROSOLS
IN THE LOS ANGELES AIR BASIN
5. REPORT DATE
May 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
A.H. Moskowitz
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
California Institute of Technology
Pasadena, California
10. PROGRAM ELEMENT NO.
1AD712 (AE-10)
11. CONTRACT/GRANT NO.
802160
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTF, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Interim 10/75-12/76
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The atmospheric aerosol was sampled with a. low pressure impactor at a
coastal, an urban, and an agricultural site in the Los Angeles air basin.
The material collected on each stage was analyzed for nitrate by direct
vaporization into a chemiluminescent analyzer, sensitive at nanogram levels.
The method responds to inorganic nitrate compounds which vaporize or decompose
below about 1200°C. Ther coastal nitrate size distribution consists mainly
of particles which have diameters greater than 2.0 ym.,. whereas the nitrate
in the agricultural region is found primarily in the submicron range... The
urban location, exhibiting characteristics of both coastal and agricultural
regions, was bimodal about the 1-2 ym range. It is believed that the submicron
aerosol is ammonium nitrate while the larger size fraction is sodium nitrate-
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
*Air. pollution
*Aerosols
*Particles
*Nitrate minerals
*Particle size distribution
Sampling
^Chemical analysis *Chemiluminescence
Los Angeles air basin
13B
07D
08G
14B
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report}
UNCLASSIFIED
21. NO. OF PAGES
42
20. SECURITY CLASS (This page)
UNCALSSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
------- |