THIRD INTERAGENCY SYMPOSIUM
ON AIR MONITORING QUALITY ASSURANCE
May 18 and 19,1977
MARRIOTT INN
BERKELEY, CALIFORNIA
SYMPOSIUM CHAIRMAN: J.J.WESOLOWSKI
CO-CHAIRMAN: W.JOHN
SYMPOSIUM CO-ORDINATORS: E. JEUNG
C. NORIYE
SPONSORED BY
CALIFORNIA AIR RESOURCES BOARD
CALIFORNIA DEPARTMENT OF HEALTH
AIR AND INDUSTRIAL HYGIENE LABORATORY
part by an EPA Training Grant
-------
AGENDA
Third Interagency Symposium
DUALITY ASSURANCE IN PARTICULATE MASTER MONITORING
May 18-19, 1977
Marriott Inn, Berkeley, California
Symposium Chairman - Jerpme J. Wesolowski
Vice-Chairman - Walter John
Coordinators - Edward Jeung, Carol Noriye
May 18
8:30 - 9:00 Registration
9:00 - 9:15 Welcome: J. J. Wesolowski,
California Department of Health
9:15 - 9:^5 Keynote Speaker: M. Nichols, Vice-Chairperson
California Air Resources Board
9:1*5 - 10:00 Break
SESSION I PARTICUIATE MATTER AND ITS EFFECTS - Moderator: A. Gordon,
California Air Resources Board
10:00 - 10:30 Current Status of Particulate Matter Monitoring -
W. John,
California Department of Health
10:30 - 11:00 Health Hazards of Particles - J. Goldsmith,
California Department of Health
11:00 - 11:30 Mutagenicity of Particulate Matter - P. Flessel,
California Department of Health
11:30 - 12:00 Visibility Reduction - R. Charlson,
University of Washington
12:00 - 2:00 Lunch
-------
SESSION II
TECHNIQUES FOR SIZE-SELECTIVE SAMPLING
Moderator: W. Kirchhoff
National Bureau of Standards
2:00 - 2:30 Respirable Fraction - T. Mercer,,
j University of Rochester
I
2:30 - 3:00 Cyclones - G. Reischl,
California Department of Health
3:00 - 3:30 Impactors - K. Rao,
Stanford Research Institute
3:30 - U:00 Virtual Dichotomous Sampler - R. Stevens,
Environmental Protection Agency, RTP
V:00 - ^:30 Sequential Filtration Sampler - R^ Eldred?
: University of California, Davis
J
U:30 - 5:00 Standing Air Monitoring Work Group Strategy for
Particulate Monitoring - S. Sleva,
Environmental Protection Agency, RTP
6:30 - 7:00 Social Hour
7:00 - Dinner - Speaker: Harold Johnston,
University of California, Berkeley
SESSION III
May 19
PROBLEMS IN QUALITY ASSURANCE FROM PRESENT PARTICULATE
MONITORING PRACTICES - Moderator: S. Duckworth,
California Air Resources Board
,9:00 - 9:30
9:30 - 10:00
("•
10:00 - 10:30
10:30 - 11:00
11:00 - 11:30
11:30 - 12:00
Canadian Experience with Particulate Sampling, K. Miller,
Environment, Canada
Operational Problems - J. Wendt,
California Air Resources Board
Break
Composition of Particulate Matter in California
Hi-Vol Samples - S. Twiss, i
California Department of Health
Fugitive Dust - J. McCauley,
U.S. Geological Survey
Size Distribution Study of Particulate Matter in
Los Angeles and Anaheim - S. Witz,
South Coast AQMD
12:00 - 2:00
Lunch
-------
SESSION IV DISCUSSION OF PARTICUIATE MONITORING STRATEGY -
Moderator: C. Eller
Environmental Protection Agency, Region IX
Panel Discussion with Audience Participation - Panel
Members from Various Agencies
2:00 - 3:00 Panel I
What Should We Do to Improve Present Particle
Monitoring Practices?
S. Witz South Coast AQMD
M. Foley San Diego APCD
S. Sleva Environmental Protection Agency
W. Siu Bay Area Air Pollution Control District
B. Appel California Department of Health
3:00 - l4-:00 Panel II
Should We Have a Size-Selective Particle Standard?
S. Ryan South Coast' AQMD
R. Cummins Environmental Protection Agency, Region IX
J. Kinosian California Air Resources Board
K. Miller Environment, Canada
D. Johnson Texas Air Pollution Control Board
Closing Remarks: J. J. Wesolowski,
California Department of Health
-------
PREFACE
The Third Interagency Symposium on Quality Assurance focused on particulate
matter monitoring. The symposium was structured to discuss possible answers
to two major issues: l) is it necessary and feasible to have a particle
standard based on size-segregation, and 2) how can we improve the quality
of the data from the present hi-vol method of monitoring. During the first
three sessions there were talks by technical experts on particulate matter
and its effects, techniques for size-segregation sampling and problems in
quality assurance from present particulate monitoring practices. This
volume contains the contributions from many of the speakers. Note that
the speakers come from a broad range of disciplines, for example, medicine,
chemistry, physics, and geology. This demonstrates that the solution to
this particular problem requires cooperative input from different fields
of expertise.
Although the chemical characteristics of particles are related to the size-
segregation problem, no formal presentations on the chemistry of particles
were included, simply because the limited time available made it necessary
to concentrate all efforts on the physical aspects of particle sampling and
size-segregation. However, a topic related to biochemistry was introduced,
viz., the use of a bacterial test for determining the mutagenicity of ambient
air- The use of this test is relatively new and we felt that its potential,
as well as the care that must be used to obtain meaningful data, should be
brought out at this symposium.
In the past, I have heard the comment made that the reason we cannot consider
particle sizing is that we do not have good methods. This statement cannot
be made today, as I trust the following papers will attest. Yet it is true
-------
that no system'can be recommended with a guarantee that it will serve the
monitoring community well for many years. This does not mean we should
not begin at least a limited size-segregation monitoring effort while con-
tinuing with the present hi-vol system. Although laboratory testing is
absolutely essential, the final determination of the value of a monitoring
method for size-segregation must be made in the field. We did not refrain
from measuring the gaseous pollutants, such as 03, S02, WO , until we had
JC
methods guaranteed to give excellent results. As you know, some of the
methods have been improved over the years and some even replaced. There
seems to be a stigma attached to admitting a method needs improvement or
replacement. What nonsense!
Another reason given for not moving forward with size-segregation is that
the equivalency between the hi-vol method and the size-segregating method
of choice must be determined first. A January 1977 report on "Strategies
for Improved Suspended Particulate Monitoring" prepared by EPA's Standing
Air Monitoring Work Group (SAMWG), clearly points1 out the many deficiencies
of the hi-vol method, including such fundamental problems as the fact that
the collection efficiency for the larger particles is heavily influenced
by wind speed, wind direction, and the lack of differentiation between dif-
ferent size particles. Yet, the report goes on to state: "No practical
procedures for determining equivalency of alternative methods to the par-
ticulate FRM (Federal Reference Method) currently exist or are anticipated
in the near future. As a result, the hi-volume sampler is the only cur-
rently acceptable method. Considering the above, SAMWG believes the future
monitoring strategy for particulate matter should include the continuation
of the use of the hi-volume sampler. No major immediate effort should be
made to replace the current FRM; however, studies to improve and clarify the
— 2 —
-------
existing FRM are strongly suggested. Exploratory studies to evaluate
\
the long-range possibility of replacing the hi-vol are also encouraged."
I am a strong proponent of demonstrating equivalency before replacing a
method. However, there are times when an old method, such as the hi-vol,
has such severe shortcomings that we should not preclude the possibility of
changing to a new method even if equivalency cannot be demonstrated. Natu-
rally, I expect a change will take time and that parallel sampling will have
to be carried out for years.
There is another reason often given for not moving ahead with size-segregation
pilot programs and that is the cost. There is no doubt that data costs money.
Moreover, an adequate percentage of the cost must be allocated for quality
assurance. A National Academy of Sciences report entitled "Early Action on
the Global Environmental Monitoring System", states: "The quality control
of data to be used is critical to the success of the system. All data should
have associated with them measures of accuracy and precision. This will cost
time and money. Perhaps 25$ of total monitoring costs is a realistic estimate."
Further, a December 21,.1976, EPA draft report entitled "Proposed Air Monitor-
ing Strategy for State Implementation Plans" states: "The proportion of
monitoring resources used in quality assurance is inadequate (approximately
6%) and should be increased over the next five years to approximately 10 to
20% depending on the type of monitoring program being implemented." I estimate
that the quality assurance aspect of a size-segregation monitoring program will
be even greater than 25$ at first, although it should decrease as the methods
are improved.
3 -
-------
Nevertheless, we must place the quality of data very high on our pribrity list
for a number of reasons. One of the most obvious is that important and expen-
sive control decisions are based on monitoring data, and are often challenged
in the courts. A second is a more subtle one. The science of air pollution,
although by no means new, has developed substantially only over the past
few decades. I believe the heart of this science lies in the data generated
by measurements. For other physical sciences such as physics, chemistry,
astronomy, we place emphasis on the quality, rather than on the quantity of
the measurements, and centuries of tradition have made this emphasis into a
integral characteristic of these disciplines. The traditions of the science
of air pollution are just beginning to be developed now. I trust the impor-
tance of quality data will be part of that tradition lest this important
science be relegated to a second class citizenship among the physical sciencest
The truly bright and competent youngsters in our universities and colleges
do not usually choose to spend their life working in a field which is con-
sidered to be second class. But, a monitoring system is only as good as
the people who run it.
J. J. Wesolowski
-------
THE CURRENT STATUS OF PARTICULATE MATTER MONITORING*
Walter John, Ph.D.
California Department of Health
Berkeley, California
For presentation at the Third Interagency Symposium on Air Monitoring
Quality Assurance, Marriott Inn, Berkeley, California, May 18 and 19, 1977
-------
THE CURRENT STATUS OF PARTICULATE MATTER MONITORING
Walter John, Ph.D.
California Department of Health
Introduction
Particulate matter monitoring practices have been slow to change. The basic
sampler, the hi-vol, has been employed in California for a quarter of a
century. Some improvements have been made, such as the use of flow controllers,
and analytical chemistry techniques have been refined. More recently, a new
glass fiber medium has been introduced in an effort to reduce the artifactual
sulfate formation. But the hi-vol sampler remains basically unchanged. Even
if its many defects were eliminated, the hi-vol would still furnish only a
total particulate matter sample. Meanwhile it is becoming increasingly apparent
that the employment of size-selective sampling should be given serious consi-
deration.
Decisions concerning monitoring techniques must, of course, be made in light
of the goals of monitoring, i.e., the ultimate use of the data. The paramount
reason for monitoring is the protection of the public against respiratory
hazards. Particulate matter represents a large class of air pollutants
including sulfates, nitrates, metals, organics, minerals and biological
material, emitted directly by sources and formed by chemical reactions in the
atmosphere. Many of these materials are known to be toxicj others are sus-
pected toxins. Thus, we have United States and California standards on the
total particulate mass as protection against non-specific particle hazards as
well as California standards on specific species such as lead and sulfate. It
is likely that the list of specific material standards will grow as the medical
justification for such standards develops. Carcinogens have been identified
in particulate matter; the problem is to quantify the carcinogenicity. The
-------
newly developed tests for mutagenicity such as the Ames test, may provide
practical means for routine testing, and possibly, the basis of a new standard.
Another adverse effect of suspended particulate matter is the reduction of
visibility. Visibility is one of the layman's main yardsticks of air quality.
It is undeniable that good visibility contributes importantly to the quality
of life and is, therefore, one of the legitimate goals of air monitoring.
Similarly, it is desirable to minimize soiling which is not only a nuisance
but incurs economic loss as well.
Hie desirability of size-selective sampling stems from the fact that the
deposition of particles in the lungs and the reduction of visibility by light
scattering depend strongly on particle size. Samplers which can segregate
particles by size have been available for some time. In the last few years,
the development of these samplers has accelerated in response to the increased
interest in them.
Recent reassessment of particulate monitoring strategy by the U.S. Environmental
protection Agency, the sponsorship by the EPA of the development of a practical
version of the virtual impactor, and the increasing work on size-selective
samplers in California indicate that we may be at a turning point in particu-
late monitoring. It appears likely that new sampling techniques and concomi-
tant new particle standards will be introduced. Efficient solutions to the
4.
many problems posed by these developments will require a multidisciplinary
attack and judicious choices from among the many available options.
- 2 -
-------
Sampling Farticulate Matter
Probably the weakest link in the measurement of the concentration of particulate
matter in the atmosphere is the sample collection, which employs relatively
crude methods compared to the sophistication of the subsequent chemical analysis.
Sampling involves the physics of aerosols - suspensions of particles in gases.
1 2
One or more of the deposition mechanisms illustrated in Figure 1 is utilized. '
Sedimentation is the vertical settling of particles due to gravity; it is
significant only for relatively large particles, i.e., greater than about one
micrometer in diameter. Impaction results from the inability of particles to
follow streamlines because of inertia and hence, is also mainly important for
large particles. Diffusion, on the other hand, involves the random motion of
submicron particles owing to the discrete impacts^of gaseous molecules. The
cumulative effect of this bombardment is a random particle motion. Intercep-
tion of a particle results when it is too large to pass through an opening, for
example, a filter pore. Interception is particularly important for fibers
which, by definition, are considerably greater in length than diameter. Hie
aerodynamic behavior of a fiber is determined mainly by its diameter. Finally,
a particle may be deposited as a result of the action of electrical fields on
its charge.
The Hi-Vol
As the standard sampler, the hi-vol has supplied most of the available data
on particulate matter concentrations. Its long history of usage can probably
be attributed to its simplicity and low cost. However, a number of drawbacks
have been identified:
- 3 -
-------
1. The sampling efficiency depends on wind speed and direction. This arises
from the "doghouse" cover. The directional dependence could be eliminated
by adopting a cylindrically symmetric design. Dependence on wind speed is
a more difficult problem, one which has not been sufficiently investigated
for most particle samplers.
2. The flow rate affects the efficiency. This may be partially an effect of
large particle cutoff and is another aspect of the inlet configuration.
The flow rate could also affect artifactual sulfate formation and the
collection of volatile organics. These problems appear to need more
attention.
3. The glass fiber medium, generally employed because of its low pressure
drop, can give rise to artifactual sulfate and nitrate formation. The
hi-vol does not permit the use of membrane filters which are preferred
for elemental analysis by x-ray fluorescence and for minimizing artifact
formation. Consideration should be given to the adoption of a smaller
total filter sampler, a "low-vol", which would permit use of a variety of
filter media.
Since these problems affect the quality of the particulate data, it is important
that they be addressed and necessary corrective measures taken.
Size-Selective Sampling
Even if the defects of the hi-vol were corrected, it could not provide size-
selective samples. The case for size-selection is illustrated by Figure 2.
-------
PARTICLE DEPOSITION
SEDIMENTATION
IMPACTION
DIFFUSION
Hh
INTERCEPTION
ELECTRICAL
Figure 1. The various mechanisms of particle deposition are illustrated.
-------
100
FINE •* 1 » COARSE
I
NASAL-P
PULMONARY
RESPIRABLE
T-BRONCHIAL
VISIBILITY REDUCTION
ATMOSPHERIC AEROSOL
NUCLEI ^ACCUMULATION
COARSE PARTICLE
0.01
100
PARTICLE DIAMETER,
Fig-ore 2
The dependence on particle size of variables important to air monitoring are
shown.
Top: Percentage deposited in the various sections of the respiratory system.
Middle: Relative light scattering per unit mass resulting in visibility reduction.
Bottom: Volume distribution of atmospheric aerosol showing three modes. A
division at 3.5 jum into fine and coarse particle fractions is illustrated.
-------
The major adverse effects of particulate matter, which make monitoring neces-
sary, are strongly dependent on particle size. The top section of Figure 2
shows the percentage of deposition in the several parts of the human respira-
tory system taken from the work of the Task Group on Lung Dynamics: The
pulmonary (deep, non-ciliated portion of the lung), the tracheo-bronchial
(ciliated airways), and the nasopharynx (upper respiratory tract). The dashed
line is the respirable fraction as defined by the American Conference of
Governmental Industrial Hygienists.
The middle section of Figure 2 shows the relative amount of light scattering
per unit mass, showing that the predominant reduction of visibility is due to
particles smaller than ca. 2 jum. The bottom section of Figure 2 is the particle
f> 7
volume distribution in ambient air reported by Whitby and Cantrell. ' The
nuclei mode is attributed to combustion sources and is short-lived, the particles
growing by coagulation into the accumulation mode, which also includes the
products of gas-to-particle conversion. The coarse mode consists of mechan-
ically generated particles such as wind-blown dust. Although some generality
has been suggested for the trimodal distribution, the ambient particle size
distribution varies considerably with time and location. Also, while it appears
that the nuclei and accumulation modes are produced predominantly by anthropo-
genic sources and the coarse particle mode by natural sources, this is an
oversimplification. Particle sizing alone cannot be used to reliably identify
the source of the aerosol.
It is necessary to call attention to the fact that the particle diameter on
the abscissa of Figure 2 is defined differently for the three sections of the
figure. Thus, the lung deposition curves refer to the aerodynamic diameter
- 5 -
-------
which is determined by the drag force exerted on the, particle when moving
relative to air. The light scattering depends on the index of refraction of
the particle material; the pertinent size is then an "optical" diameter. For
the atmospheric aerosol, the situation is more complicated since the diameter
was measured by electrical mobility (which can be converted to aerodynamic
diameter) for particles smaller than ca. 0.5 ^m and by an optical counter for
the larger particles. Since particles in ambient air vary in density, shape
and index of refraction, the choice of which diameter to use in defining size
cuts will depend on which phenomenon is deemed most important for the problem
at hand.
It can be seen in Figure 2 that particulate sampling into two size fractions -
fine and coarse - with a outpoint at approximately 3«5 jum aerodynamic diameter
would make more direct interpretation of the data possible. Thus, the fine
fraction could be associated with deposition in the deep lung, with visibility
reduction and with the nuclei and accumulation modes in ambient air. Che coarse
fraction could be associated with deposition in the nasopharynx region and with
the coarse particle mode in ambient air.
The 3.5 jum outpoint was used for illustrative purposes. It has been frequently
chosen as a outpoint since it coincides with the 50$ point on the respirable
curve used by industrial hygienists; it is also acceptable as a boundary size
for visibility reduction and lies near the minimum between the accumulation
and coarse particle modes. It has been suggested by others that a outpoint
of 2 jim would be closer to the minimum in the ambient particle size distribution.
Another choice to be made concerns the type of particle size cutoff. A sharp
- 6 -
-------
cutoff would better separate the ambient air size modes. On the other hand,
industrial hygienists employ cyclones haying a cutoff characteristic approxi-
mating the respirable curve. It is clear that for general monitoring purposes
some compromises will have to be made in choosing particle size cuts.
Size-Selective Sampling Techniques
Given the desirability of size-selective sampling, it is useful to list the
requirements of a suitable sampler. It is emphasized that these are require-
ments for routine monitoring, not for special problems or research.
Requirements for Size-Selective Sampler for Monitoring:
1. Particles should be aerodynamically separated in order to properly assess
lung deposition.
2. Collected samples should be suitable for general chemical analysis in the
laboratory to allow detection of various chemical species.
3. "The samples should be suitable for mass determination.
h. In order to minimize the number of samples produced, a small number of
particle size cuts should be made.
5. There should be a definite cutoff at the upper end of the coarse particles,
otherwise the coarse fraction is undefined.
6. The sampling time must be specified because of loading effects. It should
be chosen to minimize costs, consistent with monitoring needs.
7. The sampling efficiency should be independent of wind direction and have
minimal dependence on wind speed.
8. The sampler should be engineered for reliable operation.
— 7 —
-------
Theoretical design of a sampler begins with solution of the Navier-Stokes
Q
equations for the air flow for the particular boundary conditions. This
involves numberical analysis on a computer. Then the particle trajectories
must be evaluated, also numerically. This procedure has been carried out
only for the cascade impactor. Even in this case, non-ideal effects are not
2 9 10
included. The design of the cyclone and virtual impactors^ is largely
empirical, based on engineering parameter studies. Likewise, filtration
theory is not very useful for practical purposes, being largely an exercise
in dimensional analysis except for the Wuclepore filter which approximates an
ideal (for the purposes of computation) geometry. Filtration theories do not
include surface effects or loading effects.
A number of available size-selective samplers are listed below with comparative
remarks:
1. inertia! impactors - The outpoint can be calculated, the cutoff can be
sharp or, in a newly-developed impactor, "respirable". Impactors are
plagued by bounce-off and reentrainmerit, however.
2. Cyclone - Cyclones have a long history in industrial hygiene, can give a
"respirable" cutoff. The cutoff can also be sharp. Reentrainment is
negligible. Cyclones are inexpensive but they do not deposit the coarse
fraction on a filter.
3. virtual impactor - The virtual impactor deposits the fine and coarse
fractions on filters with a sharp cutoff between the fractions, it is
free of bounce-off but there are some wall losses at the crossover point.
The virtual impactor is relatively expensive but the EPA is sponsoring a
project to reduce the cost.
- 8 -
-------
12
h. Special devices - These include the sequential Nuclepore filter system
and mechanical centrifuges, 3 both of which have flow rates too small for
general monitoring use.
Candidate size-selective samplers should be validated by rigorous testing
before they are fielded in a monitoring network. The tests should include the
following components:
1. Laboratory tests - Monodisperse solid and liquid particles should be used
to simulate bouncy and sticky particles in ambient air. The outpoints
should be calibrated and the sensitivity of the outpoint to flow rate
determined. Bounce-off, reentrainment and wall losses should be measured.
The performance of the sampler under various particle loadings should be
examined. The uniformity of the particle deposit should be determined.
The sampler should be examined for air leaks, especially at the filter
holders.
2. Field tests - No laboratory test can anticipate the range of variables
encountered in the field. The field tests should be conducted under a
variety of conditions with instruments to record meteorological and
other variables.
The decision to employ size-selective sampling need not be postponed until the
ideal sampler is developed; adequate samplers exist. Establishment of particle
monitoring goals will require coordinated inputs from the medical community
and the monitoring agencies. Realization of the goals will require a multi-
disciplinary scientific effort.
- 9 -
-------
REFERENCES
1. "The Mechanics of Aerosols", N. A. Fuchs, The MacMiHan Co., New York, 1961*.
2. "Aerosol Technology in Hazard Evaluation", Thomas T. Mercer, Academic Press,
New York, 1973-
3. Deposition and Retention Models for Internal Dosimetry of the Human Res-
piratory Tract, Task Group on Lung Dynamics, Health Physics 12, 173-207,
(1966). —
h. "Respirable" Dust Sampling, M. Lippmann, Amer. Ind. Hyg. Assoc. J. 31,
138-159, (1970).
5. Atmospheric Particulate Mass Measurement with Beta Attenuation Mass Monitor,
E. S. Macias and R. B. Husar, Envir. Sci. and Tech. 10, 90^-907 (1976).
6. Atmospheric Aerosols—Characteristics and Measurement, K. T. Whitby and
B. Cantrell, University of Minnesota, 1975? unpublished.
7- "Characterization of California Aerosols--!. Size Distributions of Free-
way Aerosol", K. T. Whitby, W. E. Clark, V. A. Marple, G. M. Sverdrup,
G. «J. Sem, K. Willeke, B. Y. H. Liu and D. Y. H. Pui, Atmos, Envir. £»
1+63-^82, (1975).
8. Characteristics of Laminar Jet Impactors, V. A. Marple and B. Y. H. Liu,
Envir. Sci. and Tech. 8, 6h8-6^k (1971*).
9« Dichotomous Virtual Impactors for Large-Scale Monitoring of Airborne
Particulate Matter, B. W. Loo, J. M. Jaklevic and F. S. Goulding, in
Fine Particles: Aerosol Generation, Measurement, Sampling and Analysis,
B. Y. H. Liu, Ed., Academic Press, New York, 1976, pp. 311-350.
10. Ambient Air Analysis with Dichotomous Sampler and X-ray Fluorescence
Spectrometer, T. G. Dzubay and R. K. Stevens, Envir. Sci. and Tech. 9,
663-668 (1975).
11. Collection Surfaces of Cascade Impactors, J. J. Wesolowski, W. John,
W. Devor, T. A. Cahill, P. J. Feeney, G. Wolfe and R. Flocchini, in X-ray
Fluorescence Analysis of Environmental Samples, T. G. Dzubay, Ed.,
Ann Arbor Science, Ann Arbor, Mich., 1977, pp. 121-131.
12. Analysis of Respirable Fractions in Atmospheric Particulates via Sequential
Filtration, T. A. Cahill, L. L. Ashbaugh, J. B. Barone, R. A. Eldred,
P. J. Feeney, R. G. Flocchini, C. Goodart, D. J. Shadoan and G. W. Wolfe,
Air Quality Group, Crocker Nuclear Laboratory, University of California,
Davis, 1977 (to be published).
13. Design, Performance and Applications of Spiral Duct Aerosol Centrifuges,
W. Stober, in Fine Particles, ibid.
-------
HEALTH HAZARDS
OF
ATMOSPHERIC PARTICULATE MATTER
John R. Goldsmith, M.D.
Medical Epidemiologist
Epidemiological Studies Laboratory
California State Department of Health
2151 Berkeley Way
Berkeley, California 9470*1
For Third Interagency Symposium on Quality
Assurance in Particulate Matter
May 18, 1977
-------
Sunsets, clouds, thunderstorms, and snowflakes aside, atmospheric
particulate matter affects the well-being of the human species. What
we must try to discern is how much, of what type of particulate matter,
in which physical state, and from which potentially controllable
source, shall be a matter of indifference, which shall require efforts
to abate, and which we worry about with sufficient intensity that it
should become a priority research problem.
There are (it often seems) more esoteric than practical aspects
to the matter of the health effects of atmospheric particulate matter.
In order to force this discussion toward the practical, I shall focus
on the immensely practical matter of the medical basis for air quality
standards for particulate matter.
In order to introduce some of the concepts and variables, I shall
propose three problems in schematized form.
Problem I Effect of particle size on the incorporation of lead
into the human metabolic pool.
Suppose M equals the number of milligrams of atmospheric lead
in the air breathed during a day by an average person.
Suppose these M milligrams may all be either above jfa (Stokes)
aerodynamic diameter, or all below 0.5w.
-------
-2-
What would be the relative fraction of M which would, over a
long averaging time, be absorbed and interact with the metabolic
systems of the body (or stated another way, would contribute to the
body burden of lead, if one prefers not to speculate on the metabolic
activity at different locations in the body)?
The large particulate lead would behave as though orally ingested,
and about 0.05M would be retained, the rest being excreted with the
feces.
The small particulate lead would mostly impinge onto the absorbing
or gas-exchanging surfaces of the respiratory system, and about O.kM
would be retained, the rest being excreted with the expired air.
Thus with the lead in the smaller particle size (less than O.fm).
The M milligrams collected by an air sampler would represent (in a way)
an 8 fold greater dose, than would the same amount of lead of particle
size greater than 3p«
Problem II Effect of gas-phase oxidation on particle toxicity.
Suppose M now equals the number of millimols of sulfur in the air
breathed during a day by an average person.
What is the relative toxicity of sulfur depending on whether it is
in the form of S(>2, I^SO^ or (NHif)2 SO^?
-------
-3-
The toxity, as judged by airway resistance changes in guinea pigs
for S0_ may be given by 1 M, then that for H- SO^ will be about 5M and
that for (NfiV)p SO^ about 0.15M. Note that these values are for gas,
liquid aerosol, and solid aerosol with maximum affinity for water
vapor of H_ SO^.
Problem III Effect of specific surface properties on oxidation
by carrier particles.
Suppose M equals the number of millimols of sulfur as sulfur
dioxide mixed with Nmg of one of two atmospheric particulates of 1.5u
•,' , ' ('',''„'•'" \ • , , , ;, . , ' : ^ . ' "'•;-' . . 1
with varying surface properties.. Let N be with the sample combined
Si
particulate having a smooth surface, N, be the sample mixed particles with
a highly uneven surface. What will be the relative toxicity of the two
samples?
Although quantitative data are not as abundant as for problems I &
II, the toxicity will be a function of the proportion of the sulfur
atoms oxidized, and it will be greater in N, .
So as we approach the question of what we should know in setting
scientific health protection standards for particulate matter even if we
know completely what the source is, we will need to know: particle size
state of oxidation, hydration, or ionjzation
-------
-k-
Since we don't usually know much about the source, there is a
fourth dimension, whether or not the particle i£ inherently toxic.
Lead, asbestos, arsenic, cadmium, manganese, vanadium and
freshly-fractured silica are examples of specifically toxic particu-
late matter. In general the dose to the target organ determines the
health effect, and of the three attributes listed above particle size
is the only critical one. In the case of California's air quality
standard for lead, the standard was partly derived from associations
between lead in blood and lead from high volume samplers, presumed to
reflect all except very small particles.
Occupational health standards for asbestos stipulate a numbers of
particles, of at least a given particle length, (at present 2 particles
per ml longer than 5li)»
Particle size is not consistently specified in occupational health
standards, (TLV's of ACGIH). For example the lead standard (now being
revised as to concentration) is stipulated as fumes or dust, arsenic
has no designation, cadmium has a 0.2 mg/m standard as dust or soluble
salts, but a 0.05 mg/m standard for cadmium oxide fume, manganese has
no disignation, and vanadium (V_ 0^) as dust 0.5, but as fume 0.05 mg/m .
-------
-5-
Crystalline quartz has a TLV for "respirable" dust of
, % Respirable
but for "total" dust 30mg/m for which "respirable" is defined as
% Quartz + 3
the fraction passing a size-selector with the following characteristics:
Aerodynamic Diameter (urn) % Passing Selector
unit density spere _____ _
2 90
2.5 75
3.5 50
5.0 25
10. 0
Finally amorphous silica (including natural diatomaceous earth)
has an intended TLV of 3mg/m totali or Img/m of respirable dust defined
as<5um.
Obviously not all these materials require a community air quality
standard.
A residual category of "Nuisance Particles" has a TLV of 10mg/m^ of
total dust, provided there is less than 1% quartz.
I would like to propose that we consider three different scenarios
for standards, one for specifically toxic materials, with lead as
(B)
the prototype, one for particles resulting from or involved in
atmospheric interactions, with sulfates as the prototype, and ' a class
of so-called inert particles which includes everything else. (For
convenience, let us designate these as toxic, interactive, and inert).
-------
-6-
Now, we must consider the question as to the importance of
particle size in each scenario.
A. Toxic Particles
We recognize that small (£3*5 urn) particles will contribute
greatly to uptake, as shown by deposition data in Figures 1 and 2.
Yet, the standard for the exemplary material, lead, does not specify
any particle size criterion. Why?
First, the standard is based on epidemiological data and their
convergence with experimental data, and as long as monitoring used
to judge compliance with the standard represents a similar particle
size distribution, no specification is needed. Secondly, Robinson and
Ludwig (JAPCA 17, 66^-8, 196?) studied the particle size distribution
of urban lead aerosols, and found them to be quite uniform from city
to city and sample to sample. The bulk of particles were small,
the average mass median equivalent diameter was 0.2^i and the upper
and lower quartile boundaries 0.16 and 0.^3*1. It followed that the
mass contribution of large particles was of little consequence.
Probably the tendency of aged aerosols to assume a stable size distribu-
tion function contributed to this effect, a sort of natural atmospheric
elutriation.
Should a standard be considered for asbestos or crystalline quartz,
size characteristics would need to be specified, but for other toxic
metallic particulates the same approach as with lead would be appropriate.
B. Interactive Particulates
The present sulfate standard does not specify particle size, but
neither is it based directly on epidemiological associations and their
agreement with experimental data.
-------
-7-
In the prototypical example, then, except for the presumptive
dependence on particle size characteristics of high-volume samplers,
no particle size information is applied.
Before discussing what should be used, it is desirable to define
the type of material involved in this scenario. Any particle produced
by an atmospheric reaction, or which catalyzes (or facilitates in the
sense of making more rapid) such a reaction belongs in this class. A
particle which reacts only with water vapor or ions need not be in-
cluded. However any small particle with an amorphous (rough) surface,
such as soot should be included.
Dr. Benjamin Ferris, Figure 3, has presented a figure showing the
interaction of SO,, and "suspended particulates from fossil fuels" which
was included in the so-called Rail Report (Environmental Health Perspectives
^, 97-101, 197^). As it stands, that definition is too ambiguous to fit
the "interactive" category, but I would accept the "black suspended
matter" which the British and CEC report since most of it is presumed to
be soot. The suspended particulate from fossil fuels could include fly
ash, metal fumes, cinders, and photochemical aerosols. (In fact Ferris*
figure shows data most of which were black suspended matter).
Sulfates obtained from high volume samplers (unless emitted as
sulfates, for example from a gypsum plant) would be included, as would
*
nitrates which were formed in the atmosphere.
For this group of particulates, what we need to know is not particle
size alone, but particle size, chemical nature, emission history, acidity
and oxidation state.
-------
-8-
To know only the "respirable" fraction or its mass alone, even in an
area where reactive particulate matter is expected, will not be enough.
The ideal scenario is to determine particle size-segregated mass (say^
3-5) or lu) and either physic, o-chemical characteristics (pH, specific
absorptive capacity, sulfates and nitrates with ion concentrations^or
biological activity (airway resistence, inhibition of bacterial growth,
mutagenesis, or effects on tissue culture). Such ideals are not easily
met. As a practical alternative, particle-size-segregated sulfate,
nitrate, or acidity could be determined in conjunction with epidemio-
logical studies, and could provide data suitable for standard setting,
and such studies should have high priority.
How many size cuts and at what size levels? Is sharp cutoff
necessary? There is disagreement. My personal view is that so long as
it is accurately described and reproducible any cut in the 0.5-3.5M
range would be agreeable, but validity of the characterization is more
important then any specific cutting point.
C. Inert Particles
Although we now have a standard for total suspended particulate
matter, the medical rationale is based on particle reactivity, and not
on mass alone, respirable or not.
It would follow from this that no medical rationale exists for non-
reactive non-toxic particulate matter. While visibility and soiling
criteria are quite appropriate and adequate as a basis for air quality
standards, no medically supportable basis can be discerned for standards
for inert particulate matter.
-------
-9-
I conclude that if and when criteria for reactive particles are
adequate, standards based on them should supplant total particulate
matter standards for health protection.
We should pursue data on specifically toxic substances, treating
particle size on an ad hoc basis.
The major priorities should be on measurement by at least two
classes, of particle size, character, and other activity related
parameters for reactive particulate matter, in close conjunction with
epidemiclogical studies. The choice of size classes and methods should
depend on stability and validity of both size segregation and physico--
chemical characterization.
-------
0.1
0.5 LO
PARTICLE DIAMETER,
5.0
10
FIGURE 1. Theoretical curves and measured experimental values for total respiratory tract deposition of inhaled particles. Note
fairly uniform agreement on minimum deposition at 0.2-0.5 jan: (1) theoretical, Findeisen (13), 200 ml/sec, 14 respirations/min;
(2) theoretical, Landahl (11), 300 ml/sec, 5 respirations/min, tidal volume 450 ml; (3) theoretical, Landahl (11), 300 ml/sec, 7.5
respirations/min, tidal volume 900 ml; (4) theoretical, Landahl (11), 1000 ml/sec, 15 respirations/min, tidal volume 1500 ml; (5)
theoretical, Beeckmans (16), 5 respirations/min, tidal volume 1350ml; (6) experimental, Wilson and LaMer (33), 5.5 respirations/
min; (7) experimental, Landahl et al. (9), 7.5 respirations/min, tidal volume 900 ml; (8) experimental, Gessner et al. (35), 15
respirations/min, tidal volume 700 ml; (9) experimental, Van Wijk and Patterson (S2), 19 respirations/min, tidal volume 700 ml;
(10) experimental, Dennis (77), 13.3 respirations/min, tidal volume 720 ml; (11) experimental, Dautrebande and Walkenhorst
(IS), 10 respirations/min, tidal volume 990 ml; (12) experimental, compilation by Davies (19), 15 respirations/min, tidal volume
600ml.
Source; Stuart, Environmental Health Perspectives l6;*»l-53t 1976.
-------
J 1.0
g 0.8
I 0.6
M
!"
OC
5 o.z
0.1
03 LO
PARTICLE DIAMETER, |lffl
5.0
FIGURE 2 Theoretical curves and measured experimental
values lor pulmonary or alveolar region deposition of inhaled
particles: (1) theoretical, Findeisen (IS), 14 respira-
tions/min, deposition distal to terminal bronchioles; (2)
theoretical, Findeisen (13), 14 respirations/min, deposition
in alveolar ducts and alveoli; (3) theoretical, Hatch and
Hemeon (31); (4) theoretical, Landahl 1111,300 ml/sec. 5 respira-
tions/min, tidal volume 1350 ml, deposition in bronchioles
and below; (5) theoretical, Landahl fill, 15 respirations/min,
tidal volume 450, deposition below terminal broncheoles; (6)
experimental, Wilson and LaMer (S3), 5.5 respirations/min;
(7) experimental, Brown et al. (34), 15respirations/min, tidal
volume 700 ml; (8) experimental, Gessner et al. (55); (9)
calculated on the basis of selected experimental data, Davies
(19), 15 respirations/min, tidal volume 500 ml.
Source; Stuart, oj> cit.
-------
sooo
1000
I
I
5OO -
50 -
•
•
no effect
1
• 24 hr increase
1 (tooths
• Mdoys I r^
J increase , Q3
i or exocer-l
'bation | Q
jdiseose j -f
J /—.,-.
(increase symptoms
4- decrease pulm. function
i annoyance
r~.: .
50 100 5OO KXX) 5OOO
Suspended particulates from fossil fuels
FIGURE 3 Dose—response curve for sulfur dioxide. Data
of Dr. Benjamin Ferris (personal communication).
-------
MUTAGENIC ACTIVITY OF PARTICUIATE
MATTER IN CALIFORNIA HI-VOL SAMPLES*
C. Peter Flessel, Ph.D.
California Department of Health
Berkeley, California
*For presentation at the Third Interagency Symposium on Air Monitoring
Quality Assurance, Marriott Inn, Berkeley, California, May 18-19, 1977.
-------
Abstract
The Air and Industrial Hygiene Laboratory (AIHL) has undertaken an exploratory
survey of mutagenic activity in particulate matter extracted from California
Air Resources Board Air Quality Network hi-vol filters. The mutagenic assay
employs the tester bacteria Salmonella typhimurium. developed by Ames and co-
workers (l) and has been widely applied to materials including pesticides,
food additives, cosmetics, cigarette smoke, urine, blood, industrial efflu-
ents and air particulate matter (2). The feasibility of using the assay as
an air monitoring technique is currently being examined.
The Ames Salmonella Assay has the advantages of being simple and semi-
quantitative. Assays can be done in several days at a cost per analysis of
$200-$300 Per sample once a laboratory is equipped and running smoothly.
As a monitoring tool, two general applications are possible. These are the
monitoring of stationary sources and ambient air. Special applications to
workplace environments such as those found in the petrochemical industry are
also possible since enough material for an assay can be obtained from 25
cubic meters of air.
The assay measures the extent of mutagenesis in a series of tester strains
of Salmonella, especially designed to respond to different classes of muta-
genic agents. The strains carry mutations in regions of their DNA which
direct the synthesis of the amino acid histidine, and consequently cannot
grow in the absence of the amino acid. However, specific chemical changes
in the DHA of these mutants can restore their capacity to synthesize histi-
dine and permit growth on histidine-deficient media.
- 1 -
-------
In practice,, the assay consists in treating the histidine-reguiring mutants
with the test substances on agar plates containing a trace of histidine.
After incubation, plates are scored for the presence of histidine-independent
colonies, each the result of a mutation correcting the defect in histidine
biosynthesis. In many cases, an animal cell extract, obtained from rat
liver, is added to the agar to permit enzymatic activation of precursors
to active mutagens. Since the assay is at least semi-quantitative, compari-
sons can be made between a variety of environmental agents with respect to
their mutagenic and carcinogenic potential (l). For example, the specific
mutagenic activity (revertants per microgram) of air particulate material
obtained from high-volume filters in California and New York is approximately
the same as that found in tobacco smoke condensate. Some preliminary work
on the air particulate mutagenicity of California air samples currently
underway in the Air and Industrial Hygiene Laboratory will be discussed.
- 2 -
-------
LIST OF REFERENCES
1. McCann, Joyce and Ames, Bruce N. A Simple Method for Detecting
Environmental Carcinogens as Mutagens, Occupational Carcinogenesis,
N.Y. Acad. of Sci., 1976, pp 5-13-
2. Talcott, Ronald and Wei, Eddie, Brief Communication; Airborne Mutagens
Bioassayed in Salmonella typhimurium, J. National Cancer Institute,
Vol. 58, No. 2, February 1977-
-------
DOSE RESPONSE CURVE
Mutagenic Activity of Air Particulate Matter
In Strain TA 100
1600
1200 -
800 -
400
200 400 600 800 1000
ug Air Particulate Matter per plate
-------
Location
Date
Lakeport 11/2/76
Salinas 11/2/76
Sacramento 11/2/76
Fresno 11/2/76
Riverside 11/2/76
Bakersfield 11/2/76
Bakersfield 10/27/76
Bakersfield 11/8/76
Mutagenic Activity*
per micrograa, extracted in:
Acetone Benzene
TA 98 TA 100 TA 98 TA 100
0.4
0.1
0.3
0.4
0.5
0.8
0.4
0.2
0.1
0.1
0.5
0.4
0.7
0.5
0.4
0.4
1.6
1.4
0.6
0.8
0.7
0.5
1.7
1.1
0.6
0.7
1.5
Mutagenic Activity*
per cubic meter, extracted in:
Acetone Benzene
TA 98 TA 100 TA 98 TA 100
2
5
1
9
2
2
1
8
2
3
5 6
4 4
2 5
micrograms per cubic meter
Mass Organics Lead Sulfate Nitrate
39.8
127.1
148.7
269.7
88.6
329.3
169.0
264.7
2.2
4.4
9.2
23.3
7.2
26.5
10.9
19.2
0.3
0.8
2.0
7.2
4.3
5.3
2.8
4.8
2.0
8.1
4.9
8.2
4.3
19.2
13.8
17.3
0.9
9.3
13.8
11.4
4.1
47.3
11.0
33.6
* Mutagenic Activity units defined as mean number of revertants less 1 CT,
-------
CORRELATION BETWEEN MUTAGENIC ACTIVITY and:
ACETONE
Extract
TA 98
TA 100
Benzene
Extractable
Organics
0.97
0.93
Total
Particulate
Matter Lead
0.93 0.85
0.88 0.93
Sulfate
0.75
0.81
Nitrate
0.83
0.66
Extract
TA 98
TA 100
0.85
0.79
0.85
0.79
0.87
0.85
0.65
0.63
0.65
0.63
Spearman's Rank Correlation Coefficient
Significant at 95% f or o ^ 0.71
and at 99% f or £ 0.84
-------
MUTAGENIC ACTIVITY FOB ACETONE EXTRACTS
In Strain TA 98
&
>>
o
•H
a
10-1
8H
4H
2-
10 n
80 160 240 320 400
Total Particulate Matter — ug/m3
4'
5 10 15 20 25
Benzene Extractable Organics — ug/m3
-------
PART I. CURRENT CONCEPTS OF CARCINOGENESIS
A SIMPLE METHOD FOR DETECTING
ENVIRONMENTAL CARCINOGENS AS MUTAGENS *
Joyce McCannt and Bruce N. Ames
Department of Biochemistry
University of California
Berkeley, California 94720
In the past several decades there has been an extraordinary proliferation of
man-made chemicals in the environment and the working place, with literally
thousands of new chemicals added every year. It is a subject of increasing
concern that, with few exceptions, these chemicals have not been tested to
assess their potential danger to humans as mutagenic, carcinogenic, or terato-
genic agents. To test such large numbers of chemicals by use of conventional
animal systems would be virtually impossible because of economic, space, and
time limitations.
The rapidly accumulating evidence that, with few exceptions, carcinogens
are mutagens, has confirmed the desirability of using simple, rapid, and eco-
nomical test systems capable of detecting carcinogens as mutagens as a pre-
screening technique to pinpoint potentially dangerous chemicals requiring more
extensive testing in conventional animal systems. A great deal of evidence that
carcinogens are mutagens has been obtained using a very sensitive and simple
bacterial test for detecting chemical mutagens.1"8 The development of this
test system, recent improvements in the test, and evidence indicating that the
test is reliable and efficient for the detection of carcinogens as mutagens will be
summarized in this brief review.
THE BASIC TEST SYSTEM
The test method has recently been summarized in detail." Compounds are
tested on petri plates with specially constructed mutants of Salmonella typhi-
murium as tester strains. After hundreds of mutants were screened, several
tester strains were selected for sensitivity and specificity in being reverted from
a histidine requirement back to prototrophy by a variety of mutagens. Most
mutagens can be detected by spotting a small amount (usually < 1,000 /u.g) of
the chemical on a lawn of the bacterial tester strain in a petri dish. A positive
result is seen by growth of revertant bacteria around the spot (FIGURE 1).
In this way a wide range of concentrations are tested simultaneously as the
chemical diffuses into the petri plate. For quantitative results, different concen-
trations are tested individually by incorporating the mutagen into the thin agar
overlay along with the bacteria, and dose-response curves are easily obtained
(FIGURE 2). For most mutagens linear dose-response curves are obtained. At
highe? doses revertant colonies decrease due to toxicity of the chemicals.
By adding a microsomal activation system of rat (or human) liver to the
petri plates a wide variety of carcinogens can be activated to mutagens and
* Supported by ERDA Contract AT(04—3)34 P.A. 156.
t Postdoctoral Fellow, California Division of the American Cancer Society.
-------
Annals New York Academy of Sciences
FIGURE 1. The spot test. Each petri plate contains, in a thin overlay of agar, a
trace of histidine and biotin," the bacterial tester strain (in this case TA98 was used),
and plates C and D contain in addition a liver microsomal activation system from
arochlor-induced rats.1- Mutagens were applied to small paper discs, which were
then placed in the center of each petri plate: A=spontaneous revertants; B = furyl-
furamide (AF-2) (1 ^g); C = aflatoxin B, (1 ^g); D = 2-aminofluorene (10 ng).
Mutagen-induced revertants appear as a ring of colonies around each disc. (By per-
mission of the publisher of Mutation Research")
detected easily. Thus, an important aspect of mammalian metabolism can
be duplicated in an in vitro test. A large group of carcinogens—aflatoxin Bj,
benzo(a)pyrene, 2-acetylaminofluorene, and others, have been detected as
frame-shift mutagens after liver activation 4 Each activated molecule contains
a ring system capable of stacking interaction with DNA and an electrophilic
group that can react with DNA/1 8 Other groups of carcinogens have been
detected as mutagens causing base pair substitutions: y9-propiolactone, propane-
sultone, N-methyl-N'-nitro-N-nitrosoguanidine, and so on.1'2-4 Some carcino-
gens such as nitroquinoline-N-oxide cause both types of mutations.3
-------
McCann & Ames: Detecting Carcinogens as Mutagens 7
Detection of Mutagenic Metabolites in Urine
The method has been extended for the detection of mutagenic metabolites
in urine.10-" A wide variety of metabolites of drugs and other ingested com-
pounds appear in the urine, many conjugated as j8-glucuronides. The addition
of commercial y8-glucuronidase to the petri plates along with the urine, liver
homogenate, and bacteria allows detection of metabolites that are excreted in
0.1
0.3 0.5
1.0
0.2
p.g Carcinogen
FIGURE 2A. Dose-response curves obtained for a variety of carcinogens and mu-
tagens. Mutagens were incorporated directly into the agar overlay with the bacteria
(TA100 or TA1538) and, where indicated, S9-Mix.8' " (By permission of the pub-
lisher of Mutation Research.6)
-------
8
Annals New York Academy of Sciences
urine as /J-glucuronide conjugates. By this method mutagenic activity is readily
demonstrated with urine of rats administered as little as 200 fig (1.6 mg/kg) of
the carcinogen, 2-acetylaminofluorene. This method can be used for the screen-
ing of human urines in order to detect mutagenic metabolites of drugs and of
dietary components. It may also be useful for testing urinary metabolites of
drugs and food additives in experimental animals.
BenzolaJpyrene-A.S-oxide
(TA1538)
2-Aminoanthracene + S-9
(TA1538)
Benzo(a)pyrene + S-9
(TA100)
7,12-Dimethylbenz(a)anthracene
+ S-9 (TAIOO)
I i
0 5 10 15 20 25 30
fj.g Carcinogen
FIGURE 2B. See legend for FIGURE 2A. (By permission of the publisher of Mutation
Research.")
Detection of Mutagenic Activity in Complex Mixtures
The economy of the bacterial/ mammalian-microsomal assay suggests its
usefulness as a tool for rapidly obtaining information about the mutagenic and
potential carcinogenic activity of uncharacterized compounds in complex mix-
tures, and it can be used as an assay in purifying mutagenic components from
complex mixtures. A detailed study has been made of the mutagenic activity
of cigarette-smoke condensate and 12 standard smoke condensate fractions.12
-------
McCann & Ames: Detecting Carcinogens as Mutagens 9
The condensate from less than 0.01 cigarette could easily be detected. Recently,
considerable mutagenic activity has been found in many commercial oxidative-
type hair dyes,13 and several of the individual mutagenic components in the
hair dyes have been identified. We have also shown that there is considerable
mutagenic activity in soot from city air (with D. Streitwieser, unpublished).
The Bacterial Tester Strains
Standard Tester Strains
There are several standard bacterial tester strains. One strain (TA1535)
can be used to detect mutagens causing base-pair substitutions and two (TA1537
and TA1538) to detect various kinds of frame-shift mutagens.
The molecular basis of the frame-shift mutations in these strains has been
investigated. Frame-shift mutations occur by shifted pairing in repetitive se-
quences of DNA, and frame-shift mutagens can be very specific for the particu-
lar sequences they mutate. TA1538 has a repetitive —C—G—C—G—C—G—
C—G— sequence near the site of the histidine mutation,14 and is reverted
particularly well by many carcinogens, such as 2-nitrosofluorene. The other
frameshift tester strain, TA1537, appears to have a run of C's at the site of the
mutation,3 and is reverted by activated 7,12-dimethylbenzanthracene and par-
ticularly well by 9-aminoacridine. In addition to the histidine mutation, each
tester strain contains two additional mutations that greatly increase its sensitivity
to mutagens: one causes loss of the excision repair system and the other loss of
the lipopolysaccharide barrier that coats the surface of the bacteria.3
The Development of New Bacterial Tester Strains
We have recently developed two new tester strains (TA100 and TA98) by
transferring a resistance transfer factor (R factor) to our standard tester strains
TA1535 and TA1538, respectively. These new strains are extremely effective
in detecting classes of carcinogens that we previously had not detected with our
original strains, and they are much more sensitive to a number of carcinogens
we had previously detected only weakly. These carcinogens include aflatoxin Bt,
stcrigmatocystin, furylfuramide (the nitrofuran food additive AF-2) and other
nitrofuran carcinogens, acetylenic carbamates, methyl methanesulfonate, nitro-
quinoline-N-oxide, benzo(a)pyrene, 7,12-DMBA, benzyl chloride, and ace-
toxysafrole.
The effect of the resistance transfer factor on mutagenesis has important
implications for the mechanism of mutation. We believe the class of mutagens
detected by the new tester strains may be causing nicks in the DNA, and
mutagenesis is a consequence of an error-prone repair which is enhanced by
the R factor.15 G. Walker, in this laboratory, has isolated several mutant R
factors that have lost the ability to enhance mutagenesis, and has shown that
the effect of the R factor on mutagenesis does not occur in lex mutants of
E. coll, thus implicating the involvement of the error-prone repair pathway.
New endonuclease (D. Lackey, N. Spmgarn, S. Linn, G. Walker, and B. N.
Ames, unpublished) and DNA polymerase le activities have been observed in
bacteria containing R factors, and these new enzyme activities could be involved
in the effects on mutagenesis.
-------
10 Annals New York Academy of Sciences
Three additional tester strains are under development that will be useful for
the detection of special classes of mutagens: 1) a new frame-shift tester strain
to replace TA1537; 2) a new tester strain with normal excision repair designed
to detect certain DNA cross-linking agents such as mitomycin C 15 (with E.
Choi); and 3) new tester strains lacking nitroreductase enzymes. Many nitro-
carcinogens, such as nitrofurazone and furylfuramide, are activated directly to
mutagens by bacterial nitroreductases. Mammalian nitroreductases can also
activate this class of carcinogens. Bacterial nitroreductase mutants have been
isolated by others,1*'1S and in collaboration with M. Vore we are isolating
nitroreductase mutants in our tester strains. These new strains will be useful
for comparisons between bacterial and mammalian nitroreductase activities.
Validation of the Test System
We propose that the 5a/mowe//a/microsome test be used for the screening of
food additives, drugs, and chemicals to which humans are exposed, and for the
routine screening of all new chemicals under development that are potential
sources of human exposure. The system is very inexpensive, sensitive (ng or /ig
can be detected), and rapid (results are seen in two days).
Many carcinogens and noncarcinogens have been tested with the original
four bacterial strains. We are currently compiling results obtained using the
test in this and in many other laboratories throughout the world (McCann, et al.,
in preparation). So far, about 85% of the carcinogens tested (117/139) have
been detected as mutagens. These include a wide variety of carcinogens such
as direct alkylating agents, nitrosamines, polycyclic hydrocarbons, fungal toxins,
aromatic amines, nitrofuran carcinogens, a variety of antineoplastic agents, and
antibiotic carcinogens such as adriamycin, daunomycin, and mitomycin C. Also,
most of the known human chemical carcinogens that have been tested are posi-
tive. These include /3-naphthylamine, benzidine, cigarette-smoke condensates,
BCME, aflatoxin B^ vinyl chloride, 4-aminobiphenyl, and a variety of coal-tar
components. Carcinogens not yet detected include acetamide, dimethyl hydra-
zine, safrole, urethan; metal carcinogens, such as lead acetate and titanocene
dichloride; and promoter carcinogens such as phorbol. Most of these chemicals
have not yet been tested in our new tester strains (TA100 and 98), and we feel
that the increased efficiency of these new strains in detecting carcinogens will
lend even more support to the correlation of carcinogenicity and mutagenicity.
The test is highly selective for the detection of carcinogens. To date, 59
noncarcinogens (chemicals negative in tests for carcinogenicity) have been
tested, most of which are close relatives of carcinogens. Extremely few (5/59)
are mutagenic in the test and for several of these carcinogenicity tests have been
,extremely limited, and there is some doubt as to the classification of these
chemicals as noncarcinogens. Many hundreds of common chemicals of unknown
carcinogenicity have also been tested, and in general, very few chemicals are
positive. We feel that chemicals to which humans are exposed that are clearly
positive in the test should be considered potential human health hazards, and
should be thoroughly tested in animal systems, and, where extensive human
exposure has occurred, appropriate epidemiologic studies should be done.
-------
McCann & Ames: Detecting Carcinogens as Mutagens 11
Predictive Value of Mutagenicity Testing
There are already several cases, at least one involving extensive human
exposure, where chemicals initially detected as mutagens have been subsequently
found to be carcinogens. Certainly the most dramatic of these concerns the food
additive furylfuramide (AF-2), which was used extensively in Japan from 1965
until recently as an antibacterial food additive in a wide variety of common
food products such as soybean curd and fish sausage. It was found negative in
tests for carcinogenicity in rats in 1962 and 1964,19'20 and again in mice in
1971.21 In 1973 it was found to be highly mutagenic in bacteria using a strain
of E. colt22'23 (it is highly mutagenic on TA100;1B see FIGURE 2). In fact,
the level of mutagenic activity of this chemical in food was such that one could
easily demonstrate the mutagenicity of a slice of fish sausage (T. Sugimura,
personal communication). It was subsequently examined in other eukaryotic
organisms and found to be mutagenic in yeast and Neurospora?* and to cause
chromosome breaks in human lymphocytes.25 Animal tests for carcinogenicity,
more extensive than had been conducted previously, were initiated and these
tests have recently shown that AF-2 is, in fact, a carcinogen (Ikeda, et al.,
cited in Ref. 26). As a consequence of this finding, the Japanese govern-
ment discontinued use of AF-2 as a food additive and all products already
containing AF-2 were removed from the market. Since AF-2 had already been
tested for carcinogenicity in two animal systems and found negative, it is
unlikely that further tests would have been conducted in the near future. Any
deleterious effects of AF-2 on the Japanese population would not have been
evident for decades, and it is possible that a catastrophe may have been avoided
by the early detection of this carcinogen by a simple bacterial mutagenesis test.
As it is, the Japanese people have consumed relatively large amounts of this
carcinogen for eight years. It is too early to tell what the consequences of this
exposure will be.
Another example of a carcinogen initially detected as a mutagen is dibromo-
ethane (ethylene dibromide), a widely used industrial chemical and gasoline
additive, which was detected as a mutagen more than four years ago in the
Salmonella test1 and in other systems. It was recently tested for carcinogenicity
and found positive.27
Although in general most chemicals tested are negative, there are a few
examples of chemicals of unknown carcinogenicity that are potent mutagens
in the Salmonella/ microsome test, and that are sources of human exposure which
we feel should be thoroughly tested in animal systems. Studies in this laboratory
have recently shown the mutagenicity of most common hair dyes.13 Eighty-nine
percent (150/169) of commercial oxidative-type (hydrogen peroxide) hair-dye
formulations were mutagenic, and of the eighteen components of these hair dyes,
nine were mutagenic. Hair dyes are known to be absorbed through the skin, yet
extremely few of the hair dyes or components have ever been tested for car-
cinogenicity. Such tests are now under way by the National Cancer Institute
and the hair-dye industry. Twenty million people (mostly women) dye their
hair in the United States, and the hazard could be considerable if these chemicals
are actually mutagenic and carcinogenic in humans.
We have also examined several large-volume industrial chemicals and their
possible metabolic products for mutagenic activity in the Salmonella test, and
have found that chloroacetaldehyde, a possible metabolic product of the human
carcinogen vinyl chloride, is a potent mutagen.28 Chloroacetaldehyde is also
-------
12 Annals New York Academy of Sciences
a likely metabolic product of 2-chloroethanol, a common industrial chemical,
and a possible carcinogen.20 There is considerable evidence to suggest that
1,2-dichloroethane (ethylene dichloride) (produced in excess of 8 billion
Ibs/yr), an industrial precursor of vinyl chloride and a lead scavenging agent
in gasoline, is metabolized to chloroethanol and chloroacetaldehyde. The
mutagenicity of chloroacetaldehyde, the similarity in chemical structure between
dichloroethane and dibromoethane, a known carcinogen, and the huge amounts
of dichloroethane produced annually indicate that comprehensive cancer tests
should be conducted. (The National Cancer Institute is currently testing
dichloroethane.)
Ethidium bromide, a known DNA intercalating agent, is not directly muta-
genic in our test, but after in vitro metabolic activation with rat liver homoge-
nates, it is quite active toward the TA1538 tester strain (in preparation). Ethid-
ium bromide is a common laboratory chemical used extensively in physicochemi-
cal studies with nucleic acids. We believe it should be handled with caution.
We are living in a sea of chemicals that have not been tested for their ability
to cause cancer or birth defects. The use of short-term, efficient test systems
such as the Salmonella/micTosome test described here can help to pinpoint,
among those thousands of untested chemicals to which humans are exposed,
those which should be given priority in more extensive, long-term animal tests.
ACKNOWLEDGMENT
We thank Edith Yamasaki for the dose-response curve of benzo(a)pyrene-
4,5-oxide.
[NOTE ADDED IN PROOF; Since this paper was presented, the compilation of data
obtained using the test system for detection of carcinogens as mutagens has been com-
pleted (McCann, J., E. Choi, E. Yamasaki & B. N. Ames. 1975. Proc. Nat. Acad.
Sci. USA 72. In press.]
REFERENCES
1. AMES, B. N. 1971. In Chemical Mutagens: Principles and Methods for their
Detection. E. A. Hollaender, Ed. Vol. 1: 267-282. Plenum Press. New York,
N.Y.
2. AMES, B. N. 1972. In Mutagenic Effects of Environmental Contaminants. M. E.
Sutton & M. I. Harris, Eds. : 57-66. Academic Press. New York, N.Y.
3. AMES, B. N., F. D. LEE & W. E. DURSTON. 1973. Proc. Nat. Acad. Sci. USA
70: 782-786.
4. AMES, B, N., W. E. DURSTON, E. YAMASAKI & F. D. LEE. 1973. Proc. Nat
Acad. Sci. USA 70: 2281-2285,
5. AMES, B. N., P. SIMS & P. L. GROVER. 1972. Science 176: 47-49.
6. AMES, B. N., E. G. GURNEY, J. A. MILLER & H. BARTSCH. 1972. Proc. Nat.
Acad. Sci. USA 69: 3128-3132.
7. AMES, B. N. & H. J. WHITFIELD, JR. 1966. Cold Spring Harbor Symp. Quant.
Biol. 31: 221-225.
8. CREECH, H. J., R. D. PRESTON, R. M. PECK, A. P. O'CONNELL & B. N. AMES.
1972. J. Med. Chem. 18: 739-746.
9 AMES, B. N., J, McCANN & E. YAMASAKI. 1975. Mint. Res. 31. In press.
10 DURSTON, W. E. & B. N. AMES. 1974. Proc. Nat. Acad. Sci. USA 71: 737-741.
11 COMMONER, B., A. J. VITHAYATHIL & J. I. HENRY. 1974. Nature 249: 850-852.
-------
McCann & Ames: Detecting Carcinogens as Mutagens 13
12. KIER, L. D., E. YAMASAKI & B. N. AMES. 1974. Proc. Nat. Acad. Sci. USA 71:
4159-4163.
13. AMES, B. N., H. O. KAMMEN & E. YAMASAKI. 1975. Proc. Nat. Acad. Sci.
USA. 72:2423-2427.
14. ISONO, K. & J. YOURNO. 1974. Proc. Nat. Acad. Sci. USA 71: 1612-1617.
15. McCANN, J., N. E. SPINOARN, J. KOBORI & B. N. AMES. 1975. Proc. Nat. Acad.
Sci. USA 72: 979-98 3.
16. MACPHEE, D. G. 1974. Nature 251: 432-434.
17. KADA, T. 1975. Mut. Res. 31: 263.
18. McCALLA, D. R. & D. VOUTSINOS. 1974. Mut. Res. 26: 3-16.
19. Also, K., M. KANISAWA, T. OKAMOTO & N. AIKAWA. 1964. J. Food Hyg. Soc.
Japan 5: 120-129.
20. Aiso, K., M. KANISAWA, H. YAMAOKA, K. TATSUMI & N. AIKAWA. 1962. J.
Food Hyg. Soc. Japan 3: 365-373.
21. MIYAJI, T. 1971. Tohoku J. Expt. Med. 103: 331-369.
22. KONDO, S. & H. ICHIKAWA-RYO. 1973. Japan. J. Genet. 48: 295-300.
23. KADA, T. 1973. Japan. J. Genet. 48: 301-305.
24. ONG, T. & M. M. SHAHIN. 1974. Science 184: 1086-1087.
25. TONOMURA, A. & M. S. SASAKI. 1973. Japan. J. Genet. 48: 291.
26. TAKIZAWA, H., M. HOZUMI, T. SUGIMURA & G. T. BRYAN. 1975. J. Nat. Cane.
Inst. 54: 487.
27. POWERS, M. B., R. W. VOELKER, N. P. PAGE, E. K. WEISBUROER & H. F. KRAY-
BILL. 1975. Abstr. 14th Ann. Mtg. Soc. Toxicol. Paper no. 123: 99.
28. McCANN, J., V. SIMMON, D. STREITWIESER, & B. N. AMES. 1975. Proc. Nat.
Acad. Sci. USA 72. In press.
29. MASON, M. M., C. C. GATE & J. BAKER. 1971. Clin. Toxicol. 4: 185-204.
-------
Brief Communication: Airborne Mutagens Bioassayed in Salmonella typhimurium 1-2
Ronald Talcott and Eddie Wei3-4
ABSTRACT—Participate airborne pollutants, collected in Buf-
falo, New York, and Berkeley, California, were assayed for muta-
genic activity in the Ames Salmonella typhimurium test system.
Mutagens requiring liver enzymes for activation, as well as direct
acting mutagens, were readily detected In the Buffalo sample. By
contrast, only direct acting mutagens were detected in the Berke-
ley sample.—J Natl Cancer Inst 58: 449-451,1977.
The presence of chemical carcinogens in the panicu-
late matter of city air has been amply documented (1-3).
However, it is less clear if these air pollutants, which are
demonstrably carcinogenic to experimental animals (4-
6), are present in sufficient amounts in city air to affect
cancer risk in man. To assess the significance of air-
borne carcinogens in human cancer, it is necessary to
correlate the total carcinogenic potential of air to the
incidence of cancer in the people who breathe it. Chemi-
cal determinations of BP, as well as other PAH in air,
have been used as indexes of the carcinogenic potential
of airborne particulates (1-3), but it is uncertain what
proportion of the carcinogenicity is reflected by BP
levels. The recent development by Ames and associates
(7) of a rapid, economical, and sensitive bacterial system
for detecting mutagens and carcinogens which act as
mutagens may provide an alternate method for assess-
ing the mutagenic and carcinogenic potentials of air
pollutants. In this report, we describe the application of
the Ames system for the bioassay of mutagenic panicu-
late airborne contaminants.
The principal samples of particulate air pollutants
selected for testing were collected by Winkelstein and
his associates (8-10) in conjunction with a series' of epi-
demiologic studies on the relationship of air pollution to
human chronic disease. Over 2,500 samples were col-
lected at 21 locations in Buffalo, New York, and its
environs. This investigation, preliminary to a more
thorough follow-up of the Buffalo studies, demon-
strates two types of mutagenic activities in selected filters
of the Buffalo collection.
MATERIALS AND METHODS
The collection of 24-hour Buffalo air samples in 1961-
63 has been detailed in (9). The glass fiber filters con-
taining the deposited particulates were folded between
cardboard, sealed with wax paper, and stored in manila
envelopes prior to analysis. Under these storage condi-
tions, organic particulate matter is generally considered
to be stable (11, 12). In addition, to verify that the
observed activities were not artifacts of storage, a freshly
collected sample of airborne particulate matter was ob-
tained on a synthetic fiber filter placed in an air duct of
the Biochemistry Building, University of California,
Berkeley. The results obtained with this more recent
sample are included here. The particular Buffalo sam-
ples used for mutagenicity assays were collected on
March 14-20, 1962, at station 8 in Buffalo, New York—a
station located downwind from a steel mill. The amount
of total suspended particulates collected at this station,
ranking eighth highest among the 21 stations, averaged
126 /ig/m3.
To determine the appropriate solvent for extracting
the filters, benzene, acetone, chloroform, and methanol
were used in Soxhlet apparatuses. Equal portions of
four filters were extracted with 200 ml of each solvent
for 2 hours. The extracts were then concentrated to
dry ness, and the residues were weighed, redissolved in
dimethyl sulfoxide (10 mg residue/ml), and assayed in
the Salmonella mutagen test system developed by Ames
et al. (7). The total mutagenic activity of each extract
equaled the specific mutagenic activity (in U revertants/
100 jug) multiplied by the amount of residue obtained
from the extraction. Under these conditions, chloro-
form and acetone extracts contained the highest total
mutagenic activities, and acetone was selected for use in
subsequent extractions. Acetone extracts of blank filters
were not mutagenic.
Since airborne particulate matter contains PAH (1-3),
an experiment was designed to estimate the percent
recovery of PAH from the filters. [3H]BP (Amersham-
Searle Corp., Des Plaines, 111.) was added to 50 cm2 of
one filter prior to Soxhlet extraction. An aliquot of the
subsequent extract was tested by thin-layer chromatog-
raphy in a system that separates BP from its oxidation
products (13). Four hours of extraction with 200 ml
acetone provided quantitative recovery of the [3H]BP
with no apparent decomposition.
Based on these preliminary studies, the following pro-
cedures were adopted for the preparation of organic
extracts: a) A 100-cm2 portion of the filter was cut and
placed in a Soxhlet apparatus, b) The filter was ex-
tracted for 4 hours with 200 ml acetone, c) The extract
was concentrated to dryness in vacuo. d) The residue
was redissolved in a small volume of acetone, trans-
ferred to a preweighed vial, dried, weighed, and redis-
solved in dimethyl sulfoxide (10 mg residue/ml).
The mutagenic activities of the organic extracts were
examined in Salmonella typhimurium strains TA1535,
TA1537, TA98, and TA100 in the presence and absence
of liver s-9. The methods described by Ames et al. (7)
were followed without modification. The extracts were
ABBREVIATIONS USED: BP = benzo[a]pyrene; PAH = polycyclic aro-
matic hydrocarbons; s 9 = 9, OOOXg- supernatant; AHH = aryl
hydrocarbon hydroxylase.
'Received May 27, 1976; accepted July 2, 1976.
"Supported by Public Health Service grant CA01323 from the
National Cancer Institute.
3Department of Biomedical and Environmental Health Sciences,
School of Public Health, Berkeley, Calif. 94720.
4 We thank Dr. B. N. Ames, Department of Biochemistry, Univer-
sity of California at Berkeley, for the Berkeley air pollution material
and for the Salmonella typhimurium test strains. We also thank Dr. W.
Winkelstein, School of Public Health, University of California at
Berkeley, for the Buffalo air pollution filters.
VOL. 58, NO. 2, FEBRUARY 1977
449
J NATL CANCER INST
-------
450
TALCOTT AND WEI
mutagenic toward strains TA1537, TA98, and TA100;
consequently these strains were studied for dose-re-
sponse relationships. We investigated the effect of AHH
induction on the expression of the mutagenic activities
by substituting normal liver s-9 in place of the Aroclor-
induced preparation and also by testing 7,8-benzofla-
vone as an inhibitor of enzyme-dependent mutagenesis
(12). These latter studies were carried out in strain
TA100, the strain exhibiting the highest sensitivity to the
organic extracts.
RESULTS AND DISCUSSION
Particulate air pollution, collected on glass fiber filters
and extracted with acetone, was mutagenic toward three
S. typhimurium strains that respond to frameshift muta-
gens (text-fig. 1). The mutagenic activity was maximally
expressed in the presence of liver s-9 from rats given
Aroclor 1254 (text-fig. 2). This activity was inhibited 50%
by the addition of 7,8-benzoflavone (100 jug/plate) and
was reduced 90% by the substitution of control liver s-9
(text-fig. 2). These results suggest that most of the muta-
genic activity of the samples can be attributed to the
presence of PAH, since PAH are also activated by Aro-
clor 1254-inducible AHH, an enzyme specifically in-
hibited by benzoflavone (14). Moreover, on thin-layer
chromatography (13), the major fluorescent material in
100
200
ug extract /plate
TEXT-FIGURE 1.—Mutagenic accivity of air pollution extract. A por-
tion of a glass fiber filter containing paniculate air pollutants col-
lected in Buffalo was extracted in a Soxhlel apparatus as described
in "Materials and Methods." The extract was tested for mutagenic
activity with S. typhimurium strains TA100 (la), TA98 (Ib), and
TA1537 (Ic) by the methods of Ames et al. (7). o = liver s-9 included
in assay; A=liver s-9 omitted.
100
200
ug extract/plot!
400
TEXT-FIGURE 2.—Enzymatic activation of air pollution extract. The
effect of AHH induction on the activation of Buffalo air pollutants
was studied with S. typhimurium TA100 as the indicator strain. »=s-9
from Aroclor 1254-treated rat liver used in assay system (7); o=s-9
from untreated rats; n=s-9 from livers of rats given Aroclor in-
hibited with 100 /ug 7,8-benzoflavone/plate; A=s-9 omitted.
the extract presented the same retardation factor value
as BP (unpublished observations). Fractionation of the
acetone extracts on high-pressure liquid chromatogra-
phy (15) should indicate whether BP or other PAH
principally account for the activation-dependent muta-
genicity.
The results obtained on strains TA98 and TA1537
indicate that a second type of chemical mutagen was
present in the acetone extract. In these strains, 20-25%
of the total mutagenic activity was expressed in the
absence of liver s-9 (text-figs. Ib, Ic), which indicated
the presence of mutagens that do not require metabolic
activation.
The presence of direct-acting mutagens in the ex-
tracts of selected filters from the Buffalo collection
raised the possibility that these compounds may have
formed by oxidation during storage. However, the pres-
ence of direct-acting mutagens in Berkeley air soot had
been observed by Streitweiser and Ames in freshly col-
lected samples (Ames B: Personal communication).
Therefore, we obtained a sample of Berkeley air soot
from Ames, extracted it, and tested it (text-fig. 3). In
contrast to the Buffalo samples, the Berkeley sample
was maximally active in the absence of liver s-9, and no
BP could be detected by thin-layer chromatography. In
agreement with the observations of Streitweiser and
Ames (Ames B: Personal communication), liver s-9 ap-
peared to lower the mutagenic activity (text-fig. 3).
These results indicate that the direct-acting mutagens
present in the Buffalo samples are probably not artifacts
of storage. It is likely that the activities of the Buffalo
samples have changed somewhat during the 14-year
storage period. But, since these samples were collected
and stored under identical conditions, comparisons of
J NATL CANCER INST
VOL. 58, NO. 2, FEBRUARY 1977
-------
BIOASSAY OF AIRBORNE MUTAGENS
451
I200|-
ZOO
ug extroct /plate
400
TEXT-FIGURE 3.—Mutagenic activity of Berkeley air pollution extract.
A portion of a filter containing paniculate air pollution collected in
Berkeley was extracted as described in "Materials and Methods."
The extract was tested for mutagenic activity with S. typhimurium
strains TA100 (3a), TA98 (3b), and TA1537 (3c) by the methods of
Amesetal. (7). o=livers-9 included in assay; A=livers-9 omitted.
relative activities and correlations to epidemiologic vari-
ables should still yield meaningful data.
The chemical identities of the mutagens in air pollu-
tion extracts require further investigation, but the pres-
ence of at least two types of mutagens agrees with recent
results obtained by Gordon et al. (16) in their studies of
air pollution fractions. These workers found that the
PAH fraction and a non-PAH fraction contained cell-
transforming agents. The observation that polluted air
contains two classes of mutagens and/or carcinogens
suggests that measurements of PAH or BP content in air
may not reflect its total carcinogenic potential. For ex-
ample, the Berkeley air sample used in this study (text-
fig. 3), was mutagenic, even though BP was present only
in submutagenic quantity if at all. The results of this
study, and those of Gordon et al. (16), suggest that it
may be useful to supplement measurements of airborne
BP levels with a bioassay. The methodologic simplicity
and quantitative nature of the Salmonella mutagenicity
test system indicate that this bioassay may be applicable
to the assessment of the toxic potential of airborne
pollutants and other complex mixtures of combusted
material (17).
REFERENCES
(1) SAWICKI E: Airborne carcinogens and related compounds. Arch
Environ Health 14:46-53, 1967
(2) HENDERSON BE, GORDON RJ, MENCK H, et al: Lung cancer and
air pollution in south central Los Angeles county. Am J Epide-
miol 101:477-488, 1975
(3) PIKE MC, GORDON RJ, HENDERSON BE, et al: Air pollution. In
Persons at High Risk of Cancer (Fraumeni JF, Jr, ed.). New
York, Academic Press, 1975, pp 225-241
(4) HUEPER WC, KOTIN P, TABOR EC, et al: Carcinogenic bioassays
on air pollutants. Arch Pathol 74:89-116, 1962
(5) KOTIN P, FALK HL, MADER P, et al: Aromatic hydrocarbons. I.
Presence in the Los Angeles atmosphere and the carcinogenic-
ity of exhaust extracts. Arch Ind Hyg Occup Med 9:153-163,
1964
(6) HOFFMAN D, WYNDER EL: Chemical analysis and carcinogenic
bioassays of organic paniculate pollutants. In Air Pollution.
Analysis, Monitoring, and Surveying (Stern AC, ed.), vol 2.
New York, Academic Press, 1968, pp 187-247
(7) AMES BN, McCANN J, YAMASAKI E: Methods for detecting
carcinogens and mutagens with the Salmonella/mammalian mi-
crosome mutagenicity test. Mutat Res 31:347-364, 1975
(8) WINKELSTEIN W JR, KANTOR S: Stomach cancer. Positive associa-
tion with suspended paniculate air pollution. Arch Environ
Health 18:544-547, 1960
(9) WINKELSTEIN W JR, KANTOR S, DAVIS EW, et al: The relation-
ship between air pollution and economic status to total mortal-
ity and selected respiratory system mortality in men. I. Sus-
pended particulates. Arch Environ Health 14:162-171, 1962
(10) WINKELSTEIN W JR, KANTOR S: Prostatic cancer. Relationship to
suspended particulate air pollution. Am J Public Health
59:1134-1138, 1969
(11) ANONYMOUS: Tentative method of analysis for suspended partic-
ulate matter in the atmosphere. Health Lab Sci 7:279-286, 1970
(12) ANONYMOUS: Collection of airborne particles for analysis of poly-
cyclic organic matter. In Particulate Polycyclic Organic Matter.
Washington, D.C., National Academy of Sciences, 1972, pp
253-260
(13) KINOSHITA N, SHEARS B, GELBOIN HV: K region and non K
region metabolism of benzo[a]pyrene by rat liver microsomes.
Cancer Res 33:1937-1944, 1973
(14) FELTON JS, NEBERT DW: Mutagenesis of certain activated carcin-
ogens in vitro associated with genetically mediated increases in
monooxygenase activity and cytochrome P,-450. J Biol Chem
250:6769-6778, 1975
(15) NOVOTNY M, LEE ML, BARTLE KD: The methods for fractiona-
tion, analytical separation, and identification of polynuclear
aromatic hydrocarbons in complex mixtures. J Chromatogr Sci
12:606-612, 1974
(16) GORDON RJ, BRYAN RJ, RmMjS.etal: Transformation of rat and
mouse embryo cells by a new class of carcinogenic compounds
isolated from city air. Int J Cancer 12:223-232, 1973
(17) KIER LD, YAMASAKI E, AMES BN: Detection of mutagenic activ-
ity in cigarette smoke condensates. Proc Natl Acad Sci USA
74:4159-4163, 1974
J NATL CANCER INST
VOL. 58, NO. 2, FEBRUARY 1977
-------
MUTAGENIC ACTIVITY OF AIRBORNE PARTICIPATE ORGANIC POLLUTANTS
*
James N. Pitts, Jr., Daniel Grosjean and Thomas M. Mischke
Statewide Air Pollution Research Center, University of California,
Riverside, CA 92521
Vincent F. Simmon and Denis Poole
Microbial Genetics Program, Division of Life Sciences, Stanford
Research Institute, Menlo Park, CA 94025
Abstract
Organic extracts of airborne particulates collected at eight urban
locations and one nonurban site in Southern California were analyzed for
mutagenic activity using Ames* Salmonella typhimurium assay system. All
urban samples exhibited mutagenic activity, confirming initial findings.
The activity was observed with strains susceptible to frameshift mutation
and was not enhanced by the addition of a mammalian metabolic activation
system, thus suggesting the presence of mutagens other than benzo(a)pyrene
and other polycyclic hydrocarbons which require metabolic activation. Assay
of 0.1 to 1 mg of the samples resulted in a five- to 20-fold increase in
his revertants above the background level, with the number of revertants
per plate ranging from 0.10 to 0.44 per ug of organic carbon tested.
To whom correspondence should be addressed.
Accepted for publication in Toxicology Letters. April 22, 1977.
-------
Introduction
Of increasing concern is the possibility that chemicals released in
the environment may be a significant cause of cancer in man [1-4]. Thus,
the search for, and identification of, carcinogenic compounds in ambient
air is particularly important, as air pollutants have been implicated in
the rising rate of incidence of human cancer in urban areas [5]. Because
animal studies required to test various known or suspected carcinogens
are complex, time-consuming and expensive, experiments directed to elucida-
ting the potential carcinogenicity of air pollutants have been relatively
limited [6-8].
However, microbiological assays for mutagenic activity have recently
been developed by Ames and co-workers [9-12], who reported a correlation
of about 90% between carcinogenic and mutagenic activity (and absence
thereof) for 300 chemicals tested [11,12]. Therefore, with the clear
understanding that mutagenic activity in bacterial assays does not neces-
sarily imply occurrence of cancer in man, these relatively simple, fast and
inexpensive assays may be used as a first approach for the screening of
presumptive carcinogens [13]. Such tests appear to be especially useful
for testing complex mixtures such as polluted air samples in which several
hundreds of chemicals have been identified [14].
Thus, as part of our study of the physical and chemical transformations
of airborne pollutants in California's South Coast Air Basin, which includes
the Los Angeles area where the carcinogenic activity of air samples has been
established [6-8], we reported in 1975 [15] preliminary results showing
mutagenic activity in the organic fraction of ambient atmospheric particulates
using Ames* assay system. Subsequent to our study, mutagenic activity was
reported in ambient particulate samples collected in Ohmuta and Fukuoka,
-------
Japan [16], Buffalo, N. Y. and Berkeley, CA [17].
We report in this paper further results concerning the organic
fraction of airborne particulate matter.
Experimental
Ambient particulate samples were collected in 1976 at one nonurban
site (Camp Paivika, elevation 5,320 ft., in San Bernardino Mountains)
and eight urban locations (Anaheim, Banning, Lennox, Los Angeles, Los
Alamitos, Pasadena, Pomona and Riverside) in the California South Coast Air Basin
(Table I). Samples were collected by drawing ambient air through glass fiber
filters using high volume samplers, and the organic components were extracted by
ultrasonication with 200 ml. of equal parts by volume methanol, benzene
and dichloromethane. The organic extracts were filtered and reduced to a
volume of 1 to 3 ml. using a Kuderna-Danish concentrator, and then stored
in the dark at 5°C in glass vials with Teflon-lined caps prior to testing
for mutagenic activity.
The concentrated organic extracts were tested using the Ames Salmonella
typhimurium assay system [9-12]. Each sample was tested twice at at least
six doses (typically 0.1, 0.5, 1, 5, 10 and 50 yl) on five histidine-
dependent strains of Salmonella (TA98, TA100, TA1535, TA1537 and TA1538)
in the absence and presence of 0.5 ml. of a rat liver homogenate metabolic
activation system [18-19]. Appropriate positive and negative controls as
well as blanks consisting of the solvent mixture and of extracts obtained
from ultrasonication of the solvent mixture with unused filters and concen-
trated as indicated before were included in each assay.
In order to ensure that no artifact was created as a result of our
analytical procedure, extracts were also concentrated by vacuum distillation
of the solvent at ambient temperature. Mutagenic assay of extracts of the
same sample concentrated by the two techniques (Kuderna-Danish and vacuum
-------
4
distillation gave identical results, thus indicating that no positive
(oxidation of inactive chemical to an active one) or negative (degradation
of a mutagenic compound into an inactive form) artifacts were created when
concentrating the extract at high temperature (80°C) using the Kuderna-
Danish concentrator. Solvents recondensed after use in the concentrator
were also tested; the condensates showed no activity thus indicating that
evaporative losses of mutagenic chemicals initially present in the sample
were insignificant.
Results and Discussion
Mutagenic assays conducted on the organic extracts of airborne parti-
culates collected at nine sites in Southern California resulted in the
following findings. All solvent and filter blanks were inactive. Although
containing appreciable amounts of organic material, the sample collected
at a nonurban site (Camp Pavaika) was also inactive. All samples collected
at urban sites exhibited mutagenic activity without requiring metabolic
activation. Typically, assay of 0.1 to 1 mg of organic solvent-soluble
material (10 yl of the concentrated extract) resulted in a five- to 20-fold
increase in the number of histidine revertants per plate above the
spontaneous background in strains TA1537, TA1538 and TA98 and in up to a
two-fold increase in strain TA100.
For all active samples, the dose-response curves were found to be
linear (Figure 1) for at least the four highest doses tested (0.5, 1, 5
and 10 yl). Assay of 0.05 and 0.1 yl of some samples did not result in
a significant increase in the number of his revertants above the background
level. The number of revertants per plate per yg of organic carbon tested,
calculated from the number of revertants per plate in the linear region
-------
of the dose-response curve and the organic carbon content of the aerosol
samples (measured using an organic carbon analyzer [20]), are listed in
Table I.
For all but one urban sample which showed a slight (about 30%)
increase, addition of the metabolic activation system did not increase
the observed mutagenic activity. Strains in which the number of
revertants increased were those which are susceptible to frameshift
mutation (TA1537, TA1538, TA98 and to a lesser extent TA100). No
activity was observed with strain TA1535, which is susceptible to base-
pair substitution mutation, in any of the assays.
Although mutagenic activity of particulate organic pollutants in
Southern California urban air had not been reported prior to our initial
studies [15], the carcinogenic activity of airborne organic particulates
collected in the Los Angeles area has long been established [6-8] and
has often been attributed to the presence of benzo(a)pyrene (BaP) and
other polycyclic hydrocarbons whose concentrations in urban air have
been extensively measured [21-23]. However, since most polycyclic
hydrocarbons (including BaP) are not active in the absence of metabolic
activation in the Ames assay system, our results suggest the presence
of other active mutagens in our samples. The presence of alkylating
agents inducing base-pair substitution mutations is unlikely based on
the negative results obtained with Strain TA1535. In addition, based
on vapor pressure considerations, most alkylating agents known or suspected
to be present in polluted air (for example vinyl chloride, epoxides, nitroso-
amines) are expected to be in the gas phase rather than the particulate phase
and therefore would not be collected using our sample procedure.
-------
Other known organic constituents of airborne particulate matter include
a wide variety of primary (directly emitted into the air) and secondary
(formed by photochemical reactions involving hydrocarbons, oxides of
nitrogen, ozone and free radicals) pollutants such as long chain
alkanes, alkenes, phenols, quinones, acridines and quinolines—some
of them known mutagens [11]—, carboxylic acids, nitrate esters and
polyfunctional oxygenates [14]. Although oxygenated fractions of
ambient participates have been found to be carcinogenic [7,8], the
individual mutagenic and/or carcinogenic potency of airborne organic
oxygenates is essentially unknown. Work is in progress in our labor-
atories to identify and test those organic pollutants responsible for
the observed mutagenic activity of airborne particulate matter.
Acknowledgements
Supported by National Science Foundation-Research Applied to
National Heeds Grant No. AEN73-02904 A02 and by Stanford Research
Institute internal research and development funds. The contents of
this paper do not necessarily reflect the views and policies of the
National Science Foundation, nor does mention of trade names or commer-
cial products constitute endorsement or recommendation for use.
-------
References
[1] Heidelberger, C. (1975) Annu. Rev. Biochem., U, 79-121.
[2] Epstein, S. S. (1974) Cancer Res., 34^ 2425-2435.
[3] Searle, C. E. (1976) "Chemical Carcinogens," American Chemical Society
Monograph No. 173, Washington, D.C.
[4] Ember, L. (1975) Environ. Sci. Techno!., j), 1116-1121.
[5] Henderson, B. E., Gordon, R. J., Menck, H,, Sophoo, J., Martin, S. P.,
and Pike, M. C. (1975) Amer. J. Epidem., 101, 477-488.
[6] Kotin, P., Falk, H. L., Mader, P. and Thomas, M. (1954) Arch. Industr.
Hyg., 9_, 153-163.
[7] Wynder, E. L. and Hoffman, D. (1965) J. Air Pollut. Control Assoc., 15_,
I
155-158.
[8] Gordon, R. J., Bryan, R. J., Rhim, J. S., Demoise, C., Wolford, R. G.,
Freeman, A. E. and Huebner, R. J. (1973) Int. J. Cancer, 12, 223-231.
19] Ames, B. N., Gurney, E. G., Miller, J. A. and Bartsch, H. (1972)
Proc. Nat. Acad. Sci. USA, 6>9, 3128-3132.
[10] Ames, B. N., Lee, F. D. and Durston, W. E. (1973) Proc. Nat. Acad.
Sci. USA, 20, 782-786.
[11] McCann, J., Choi, E., Yamasaki, E. and Ames, B. N. (1975) Proc. Nat.
Acad. Sci. USA, TL> 5135-5139.
[12] McCann, J. and Ames, B. N. (1976) Proc. Nat. Acad. Sci. USA, 73,
950-954.
[13] Commoner, B. (1976) U. S. Environmental Protection Agency Report No.
EPA-60011-72-002, Washington, D.C.
[14] Grosjean, D. (1977) "Aerosols," Chap. 3, in "Ozone and Other Photo-
chemical Oxidants," Committee on Medical and Biological Effects of
Environmental Pollutants, National Research Council, Washington, D.C.
-------
8
[15] Pitts, J. N., Jr. (1975)"Chemical Transformations in Photochemical
Smog and Their Applications to Air Pollution Control Strategies,"
Second Annual Progress Report, National Science Foundation-Research
Applied to National Needs Grant No. AEN73-02904 A02, page V-8.
[16] Tokiwa, H., Tokeyoshi, H., Morita, K. , Takohashi, K., Soruta, N.
and Ohnishi, Y. (1976) Mut. Res., 38>, 351.
[17] Talcott, R. and Wei, E. (1977) J. Natl. Cancer Inst., 5£, 449-451.
[18] Kier, L. D., Yamasaki, E. and Ames, B. N. (1974) Proc. Nat. Acad. Sci.
USA, 71, 4159-4163.
[19] Ames, B. N., McCann, J. and Yamasaki, E. (1975) Mut. Res., 31, 347-364
[20] Grosjean, D. (1975) Anal. Chem., 47, 797-805.
[21] Gordon, R. J. (1976) Environ. Sci. Technol., 10, 370-373.
[22] Lao, R. C., Thomas, R. S., Oja, H. Dubois, L. (1973) Anal. Chem.,
45_, 908-915.
[23] Lee, M. L», Novotny, ~H. and Bartle, K. D. (1976) Anal. Chem., 4_8,
1566-1572.
-------
Table I. Summary of sampling conditions and mutagenic activity of airborne
organic participates.
Sampling
Site
Camp Favaika
Banning
Pomona
Riverside
Pasadena
Los Angeles
Lennox
Anaheim
Los Alamitos
Sampling
period (1976)
Oct. 4-6
Oct. 4-6
Sept. 30 & Oct. 4-6
Oct. 4-6
Sept. 29-Oct. 1
Sept. 29-Oct. 1
Sept. 25-27
Sept. 25-27
Sept. 22-24
Total
mass
collected
(grams)
0.14
0.31
0.79
0.68
0.21
0.19
0.14
0.16
0.40
TSP(a)
Pg m~
27.8
64.5
123.1
140.2
43.0
39.8
28.0
33.4
81.7
POC(b)
yg m
2.89
7.12
13.14
12.56
5.11
5.55
2.59
3.45
7.09
Number of his
revertants per plate
per yg of organic
carbon tested (c)
no activity
0.15
0.25(d)
0.19
0.30
0.44
0.24
0.21
0.10
(a) TSP « total suspended particulate levels and (b) POC = particulate organic carbon
levels, averaged over the sampling period.
(c) Data are for strain TA 1538 without metabolic activation, after substraction of
the spontaneous and solvent revertant backgrounds.
(d) This sample was slightly more active (0.34 revertants per plate per yg of
organic carbon) when tested in the presence of metabolic activation.
-------
Figure Caption
Figure 1; Dose-response curves for a sample collected at Pomona, CA
(see Table I) and tested without metabolic activation.
Each data point represents the average of two assays.
-------
.05 .1 .5 I 5 10
MIGROLITERS OF CONCENTRATED EXTRACT TESTED
Figure 1: Dose-response curves for a sample collected at Pomona, CA (see Table 1) and tested without
metabolic activation. Each data point represents the average of two assays. ^HOUT;
-------
.MEASUREMENTS OF AEROSOL OPTICAL PROPERTIES^
A. P. Waggoner
R. J. Charlson
University of Washington
Seattle, Washington 98195
^Extracted from EPA Report No. 600/9-77-001, February 1977
For presentation at the Third Interagency Symposium on Air Monitoring
Quality Assurance, Marriott Inn, Berkeley, California, May 18 and 19, 1977
-------
-JD-
MEASUREMENTS OF AEROSOL OPTICAL PROPERTIES
A. P. Waggoner
R. J. Charlson
University of Washington
Seattle, Washington 98195
ABSTRACT
Measurement of aerosol optical properties have been made in Denver
and at various rural and urban sites in California and'Missouri. Meas-
ured particle scattering coefficient has been shown to be highly corre-
lated with particle volume in the 0.1 to 1.0 ym range of particle
diameter. At times a single ionic substance (NaCl at Pt. Reyes, CA,
and H2S04/(NH4)HS04/(NH4)2S04 at Tyson, MO) controlled the aerosol op-
tics as a function of relative humidity.
INTRODUCTION
The aerosol is composed of particles that range in size from
smaller than 0.01 ym to larger than 10 ym diameter. The particles are
of various chem.ical compositions and each particle can be a mixture of
substances or a single substance. The integral optical effect of the
aerosol particles is dependent on all of these parameters. Atmospheric
optical properties normally considered would include those of interest
from a human impact standpoint, i.e., visibility and colored haze, and
those of scientific interest, i.e., scattering and absorption extinc-
tion coefficients.
Techniques have been developed at the University of Washington for
direct measurement of aerosol optical properties. These measured param*-
eters have been compared to other methods of characterizing the aerosol
impact such as visibility or particle mass loading.
-------
-36-
ATMOSPHERIC OPTICS AND VISIBILITY
It is convenient to define several parameters commonly used to
describe atmospheric optics.
The extinction coefficient bext of a real atmosphere defines the
change in intensity of light traversing a pathlength Ax by the Beer-
Lambert law:
" • - <»
b . is the sum of two terms :
ext
bext " bext (gases) + bext
bext (8ases) = bRg + bag, where
bR Ax is the fraction of incident light scattered into all direc-
tions by gas molecules in Ax.
b Ax is the fraction of incident light absorbed by gas molecules
ag in Ax.
Our interest is in b (particles) which can be broken down as
follows :
b (particles) = b + b (2)
ext ap sp ^ '
where b Ax is the fraction of incident light absorbed by particles in
ap Ax."
b Ax is the fraction of incident light scattered into all direc-
tions by particles in Ax.
The observer visibility, or visual range, is that distance at which
a black object can be just discerned against the horizon. Koschmieder1
showed that a turbid media, such as urban air, reduces the contrast
(ratio of brightness of an object to the horizon brightness, minus one)
of distant objects as given by
~bext X (Middleton2) (3)
C m C e
o
where C and C are the contrast relative to the horizon of an object
at zero°distance and at distance x. A black object has a CQ of -1.
Experiments have determined that typical observers can detect objects
on the horizon with a visual contrast of 0.02 to 0.05. Assuming hori-
zontal homogeneity of aerosol properties and illumination and a 0.02
detectable contrast, the visible range is
L - ~- (4)
v b .
ext
-------
-37-
For a contrast of 0.05,
•v - <»
ext
Usually the assumption is made that b = b
ext sp
b can be calculated from known or assumed aerosol particle size
distribution, concentration and refractive index, as discussed below.
PARTICLE OPTICS
The scattering extinction coefficient due to particles, b , Can
be calculated if the particle size distribution, number concentration
and refractive index are known and the particles are assumed to be
homogeneous spheres. None of the above assumptions are usually true
but the results of calculations show useful agreement with atmospheric
optical measurements. Figure 1 shows calculated b per volume of par-
ticle as a function of particle diameter.
The value of b is the product of the curve in figure 1 times the
particle volume distribution function. The aerosol particle volume per
log radius interval usually is similar to that of Figure 2, bimodal
with the two volume modal diameters about 0.6 urn in the 0.1 ym to 1.0
diameter range, as shown in Figure 2. In all the measurements we have
made, the particles in the 0.1 to 1.0 decade dominate Scattering extinc-
^ion in the visible spectrum although there clearly are cases in fogs,
rain, snow, clouds and dust storms in which large particles influence
or dominate visible extinction.
The correlation of bsp measured with an MRI 1550 nephelometer, and
0.1 to 1.0 urn diameter particle volume, measured using an electrostatic
mobility and single particle optical counters from Thermo Systems, was
0.95 at various locations in the Los Angeles basin. These measurements,
shown in Figure 3, are from the 1973 State of California Air Resources
Board ACHEX* program.
A correlation of bsp with the supermicrometer volume mode is not
expected unless the submicrometer and supermicrometer volume modes hap-
pen to be correlated. Thus in this qualitative sense we would not ex-
pect to find a particularly good correlation between bgp and measured
total mass concentration, for example with the high volume air sampler.
It is somewhat surprising, in view of this, that the measured cor-
relation coefficient between bsp and total aerosol mass concentration
is as high as the observed range between 0.5 and 0.9. While the former
value is not impressive nor particularly useful, the latter is suffi-
ciently high to allow inference of mass concentration from bs_. Table
1 summarizes the various published correlations of b and mass. In-
cluded in the table are correlation coefficients, r, and regression
constants A and B.
-------
-38-
SCATTERING COEFFICIENT .PER VOLUME
CM
s
"
0.01
0.1
1.0
10
PARTICLE DIAMETER, ym
Figure 1. Scattering coefficient per particle
divided by particle volume plotted as
a function of diameter. The particles
are assumed to be spheres of refractive
Index 1.50 illuminated by 550 nm light.
-------
-39-
o
UJ
«J
o
p
UJ
H
2
or.
o
POMONA 21:40
10-5-72
MEASURED
0.01
O.I
1.0
10
UJ
cr
UJ
O
Q
<
ce
o
— CALCULATED
0.01
O.I
1.0
10
PARTiaE DIAMETER,
Figure 2. Top:
Aerosol particle size distribution
measured at Pomona during 1972 State
of California Air Resources Board
ACHEX program.
Bottom: Calculated optical scattering by
particles, b , fot the measured size
distribution. The particles are assumed
to be spheres of refractive index 1.5.
-------
200
u
u
100
VOLUME. 0.1 - 1.0 ym VS. b.
TWO HOUR AVERAGES FROM
WEST COVENIA, RUBIDOUX
POMONA, DOMINGUEZ HILLS
CORRELATION COEFFICIENT * 0.948
O
OL
0
10
b$ , IN UNITS OF 10'V1
15
20
Figure 3. Plot of measured aerosol particle volume including only those of
0.1 to 1.Onm diameter versus measured b _. Measurements were
part of State of California Air Resources Board ACHEX program.
Data was supplied by Dr. Clark of North American Rockwell.
-------
TABLE 1. SUMMARY OF LIGHT SCATTERING-FILTERABLE PARTICULATE MASS 'CONCENTRATION STUDIES
^v. Mass
X. Sampling
^xMethod
Location ^X.
(Reference) ^x.
Los Angeles (4)
Oakland, CA (4)
Sacramento, CA (4)
New York, NY (5)
San Jose, CA (5)
Seattle, WA (5)
Boston, MA (6)
2.5 cm dia.
open face,
glass fiber
filter
r A B
0.83 -0.57 2.4
0.69 -0.40 1.3
0.95 0.0 2.2
— — —
— — —
0.83 -0.08 3.5
— ___ —
2.5 cm
dia.
Nuclepore
filter
r A B
— — —
— — —
— — —
0.92 -0.33 3.0
0.56 1.5 1.7
— — —
— — —
High
Vol ume
Air
Sample
r A B
0.53 -0.09 3.3
0.86 -0.61 2.4
0.93 -0.56 2.8
— — —
— — —
0.73 -0.26 3.6
0.86 0.15 2.0
Glass Fiber
filter behind
Lippmann-Harris
Cyclone
r A B
0.83 0.33 3.7
0.79 0.34 3.2
0.98 0.13 4.4
— — —
— — —
— — —
— — —
The parameters are: r = linear correlation coefficient; , ,
A and B defined by b (10'V1) = A + 10'^B (yg/nT).
-------
-42-
The correlation coefficient of 0.9 in New York City must be due to
either a correlation between the upper and lower volume (i.e., mass)
modes or an absence of the upper mode. The location at the 16th floor
of a Manhattan building suggests the latter since it was well removed
from sources of wind blown dust and other mechanically produced
particles.
In contrast, the low correlation coefficient in San Jose, CA, of
0.6 was obtained at a dusty athletic field, with the air intake at ap-
proximately 7 meters above the ground. In this case, the poor correla-
tion was likely due to a large and variable fraction of the aerosol in
the supermicrometer mode.
The wavelength dependence of bsp depends almost exclusively on
particle size distribution^. The results of measurements to date re-
garding the wavelength dependence fall into two categories. If the
wavelength dependence is described by a simple power law:
bsp~A-a (6)
where a is an experimentally determined exponent, the two categories
are:
1. Normal wavelength dependence where 0.5 j< a j< 2, with a mean
value of approximately 1.2.
2. Anomalous wavelength dependence where -1 _<_ a < 0.
The former case results in the attenuation of blue light from a
direct beam and its scattering into 4ir steradians around the scattering
volume. Of course, Rayleigh scattering always occurs simultaneously
and has a wavelength dependence that is similar:
V a x"4 <7>
As a result, blue scattered light (against a dark background) or red
transmitted light (from the sun or a bright white object) is no indi-
cation by itself of the presence of particles. Whether bsp or bRg
dominates is determined by the amount of particulate matter that is
present. In remote, clean marine locations at sea level, Porch et al.
showed that bg £ bRg at 500 nm. In continental, low altitude sites,
bap is usually larger than bRg, so that such hazes can often be assumed
to be dominated by bsp. However, clean arctic air intruding or air from
aloft subsiding into mid continent cities occasionally produce bst)
-------
-43-
is much above two, the blue haze has a significant input due to
On the other hand if the product of bgp times distance is of this mag-
nitude, then the haze is likely to be due to particles. Since
kRg,530 nm = 0.15 x 10~4nT , if bgp '^ 0, mountains should not appear
to be behind a haze if they are within 10 km. or so. They will, how-
ever, appear hazy if the distance is much more than 100 km. due to the
omnipresent scattering by gas molecules. Conversely, if such a distant
mountain is not visible at all, bsp » b^g and the haze is due to par-
ticles.
When viewing bright objects (the sun and moon, sunlit snow-capped
peaks and cumulous clouds) hazes with 1 _< a _< 2 of sufficient optical
depth cause the color to be reddened^''. The color thus produced is
remarkably similar to that observed through an optically thin layer of
802^0 so that the presence of color thus viewed is no proof of the
existence of NC^. To further complicate this issue, Husar-H has shown
that light scattered in the backward hemisphere calculated from typical
measured size distribution is enriched in the red wavelength^ also
causing the haze itself to appear reddened. In forward scatter this
same haze appears white. Charlson-^ showed that, in perhaps 20% of the
measured cases during August, 1969 in Pasadena, CA, there was enough NC>2
to influence the coloration of white objects viewed through the haze and
that in the remaining cases particles dominated the wavelength depen-
dence of total extinction (bext).
MOLECULAR COMPOSITION
The particle interaction with water, biological effects and complex
refractive index depend on the molecular composition. Therefore, it is
Important that the composition of various aerosol systems be classified,
particularly insofar as this determines the imaginary part of the re-
fractive index and hygroscopicity. Unfortunately, this is an area in
which so far very little work has been done. Rasmussen ^ suggested
that organic materials (terpenes) are a major source of atmospheric
particles, but did not quantify the work adequately for application
to optics. The reaction products of S02 with water and ammonia have
been shown to play an important part in urban and rural aerosols by
Junge^ although he did not attempt to relate quantitatively the com-
position with optical effects. We have preliminary data suggesting
that continental aerosol optics is often dominated by H2S04 and the
products of its neutralization with NHglS.lS.
The molecular nature of individual particles is a function of the
source and removal mechanisms for these particles. The most important
observable effect of composition on particle optics is the relationship
of bsp and relative humidity.
-------
-44-
BELATIVE HUMIDITY EFFECTS
The humidity effects in aerosol optics fall into three categories:
RH _< 100%: particles between and above water cloud
(including high RH hazes) ;
RH > 100%: unactivated particles in water clouds and fog;
RH > 100%: activated cloud droplets.
Our efforts have been limited to the first case and are discussed
in the following paragraphs .
Since a large fraction of submicrometer particles are hygroscopic
or deliquescent^ ^"18 the size distribution of an atmospheric aerosol
and hence its optical or climatological properties, depend largely on
relative humidities, even at RH < 50%.
First, light scattering always increases with humidity, although
for relatively hygrophobic systems the increase may be very slight up
to extremely high RH. While for most aerosols, such as ^804 droplets
the curve increases monotonically , definite inflection points due to
deliquescent salts are seen at some locations indicating the dominance
by rather pure inorganic substances such as (NH4)oSO/ or sea salt
15'1^19
The evolution of a distribution of droplets under conditions of
changing, subsaturation RH modifies the optical interactions between
radiation and particles, thus' changing the temperature of the environ-
ment of the particles and hence in turn the relative humidity. This
complex chain of events cannot be satisfactorily modelled until the
parameters which go into the models (dependence of particle growth on
chemistry-, optical properties of saturated and supersaturated droplets,
etc.) and the basic physical principles of the component processes are
understood.
A system has been designed and operated by this laboratory that
(over a period of about 120 seconds) sweeps the relative humidity of
air containing aerosol particles from 30% to 95%. Changes in particle
diameter are detected as changes in the scattering coefficient of the
aerosol particles!5,16,19.
In the midcontinent region 30 km southwest of St. Louis, this
system detected ^SO^NH^HSO^ANH^SO^ as dominate materials in the
0.1 to 1 ym decade of aerosol size. Injection of sub ppm concentra-
tions of NH3 converted the bsp(RH) response characteristic of H2S04 to
that of (NH4)2S04. The (NH^SO^ is detected by comparing the value of
relative humidity at the deliquescence point for the unknown sample
with that of laboratory-generated (NH^SO^ aerosol. 98% of the time
-------
-45-
either I^SO^ or (NH^^SO, was the dominant substance in terms of optical
effect15'16-
TECHNIQUES FOR MEASUREMENTS OF RELEVANT OPTICAL PROPERTIES
In the past several years our efforts have been focused on design
and testing of methods to measure aerosol optical properties that
directly determine aerosol radiative interactions. Methods for measure-
ment of these relevant integral aerosol optical properties, namely,
bSp(RH), and bap, are described in the following sections.
Consider a small volume of thickness dx illuminated by a parallel
beam of wavelength X and intensity I0 ^. For unpolarized light, the
intensity of light scattered into solid angle dtt at scattering angle 0
Is
(0)d '
A visibility meter using the operator's eye as a detector was devised
by Buettell and Brewer^O that geometrically performs the integration of
B^(Q) over solid angle to measure bgp^1. Ahlquist and Charlson^1 in-
creased the original instrument sensitivity by using a photomultiplier
tube to detect scattered light from a xenon flash lamp. Ahlquist
et al. improved the sensitivity, stability and dynamic range by sub-
stituting an incandescent lamp for the xenon flash lamp and detecting
the scattered light using digital photon counting techniques.
This instrument, called an integrating nephelometer , is shown in Figure
4. Modern versions of Buettell and Brewer's device have sufficient
sensitivity to be .calibrated in an absolute sense with t»Rg, the scatter-
ing coefficient of particle-free gases such as He, C02, CC12F2-
The geometric errors of the instrument have been studied by
Middleton1, Ensor and Waggoner^, Heintzenberg and Quenzel2^ and
Rabinoff and Hermanns and are estimated to be 10% or less for the
aerosol particle size distributions normally found in the atmosphere.
The modern instrument is alternately filled with ambient and par-
ticle-free air and the difference in scattered light intensity is pro-
portional to the scattering extinction coefficient due to aerosol
particles, bsp. The measured values of bsp in the atmosphere range
from lO'^nT1 at Mauna Loa Observatory to 3 x 10" ^m~ in polluted Los
Angeles (0.007 to 150 times the Rayleigh scattering coefficient at
530 nm).
The integrating nephelometer has become an accepted instrument for
measurement of aerosol scattering extinction. A series of patents have
been issued to the University of Washington based on the designs of the
-------
-46-
CLEAN AIR
PURGE
I NARROW BAND
If I /OPTICA1. FILTER
TUNGSTEN FILAMENT
LIGHT SOURCE
AEROSOL |
OUTLET , v
J [ I ^-OPAL GLASS
PHOTOTUBE
COLLIMATING DISKS
itr
AEROSOL
INLET
CLEAN AIR
PURGE
TUNGSTEN FILAMENT
LIGHT SOURCE
AEROSOL
OUTLET
[SCATTERING VOLUME
Figure 4» Diagram of nephelometer with enlarged view of the
partial shutter. Without the shutter, the instrument
Integrates the particle scattering coefficient over
* 7° to 170° to measure b . With the shutter in
. place, the instrument integrates over -v 90° to 170°
to measure b.
-------
-47-
authors of this report covering various aspects of the nephelometer.
Several hundred instruments have been produced and are in regular use
for both research and monitoring. High sensitivity, multiwavelength
instruments have been purchased by Institute filr Meteorologie, Mainz,
Germany, Air Force Cambridge Research Lab and the National Oceanographic
and Atmospheric Administration.
The draft version of \folume I of the ACHEX final report from
Rockwell International to the Air Resources Board, State of California,
recommends the integrating nephelometer for both long-term monitoring
and short-term surveillance of aerosol properties.
An optically thin aerosol layer over a dark surface increases the
albedo by scattering incident radiation backwards into space. The
albedo per unit thickness of an aerosol layer illuminated by a zenith
sun can be determined by integrating the aerosol volume scattering
function over the backward hemisphere of scattering angle. A partial
shutter, shown in Figure 4, can change the angle of integration of the
nephelometer so that the scattered light intensity is proportional to
the backward hemisphere scattering extinction coefficient b^sp due to
aerosol particles. b^Sp normally is in the range 0.1 to 0.2 times the
aerosol scattering extinction coefficient bsp.
b
_§£
The two aerosol parameters needed in simple radiative climatic
models are the particle backward hemisphere scattering coefficient,
b|jSpj and the particle absorption extinction coefficient, bap. There
are a number of ways of measuring ba , and none is entirely satisfac-
tory .
Long path extinction cannot be used because b is 10 m-1 to
10~ m~^ or smaller. Various techniques based on inverting angular 2_
scattering information have been used by Eiden ° and Grames et al. ,
etc. , but these methods require precise knowledge of the aerosol size
distribution, and contain errors of unknown size and magnitude, since
the scattering by irregular particles is calculated using Mie formulae
for spheres. The absorption coefficient of collected aerosol samples
can be estimated with low precision from measurement of the transmis-
sion of KBr pellets containing dispersed aerosol™. Lindberg and
Laude^' measured aerosol absorption by measuring the decrease of
diffuse reflectance of a white powder when a small amount of aerosol
Is dispersed in it.
All of the above methods, in our opinion, are poorly suited for
measurements in background locations. Measurement of the angular
dependence of the aerosol volume scattering function is difficult when
molecular scattering dominates. The methods of Volz and Lindberg
-------
-48-
requlre collecting an aerosol sample over several days, scraping the
sample off the collecting surface and dispersing the sample in another
media. Any treatment of the sample that alters the aerosol size dis-
tribution will alter the optical absorption coefficient '^1. A dif-
ferent technique for measurement of bap has been developed in our
laboratory that we believe is superior to those described above.
Atmospheric aerosol is collected by passing ambient air through a
Nuclepore filter. The filter consists of a 10 ym thick film of poly-
carbonate plastic with 0.4 ym holes etched through it. The holes are
etched along damage tracks from highly ionizing particles and are round
and perpendicular to the surface of the film. Individual particles with
a mean separation of several diameters are collected on the surface of
the filter. The filter and the particles are placed in an optical sys-
tem that illuminates the particles and the filter with a parallel beam
of, in this case, green light and collects both direct transmitted and
forward scattered light. The extinction or change in transmission
between a clean filter and the filter plus aerosol is assumed to be the
same as absorption by the same aerosol dispersed in a long column of
air. Knowing the volume of air passed through the filter during col-
lection of the aerosol, one can calculate the optical absorption co-
efficient due to particles, bap.
This method has been checked for accuracy using laboratory aerosols
of known (including zero) absorption coefficient and is described by
Lin et al.32o xhe disadvantages of the method center on errors in~
troduced by sample alteration that may take place during collection,
but the sample alteration is probably muck less than in the techniques
of Volz and Lindberg. The sample collection is simple and only requires
10 to 20 yg/cm of aerosol on the filter.
ATMOSPHERIC MEASUREMENTS AND DATA
fcsp and Visibility
As discussed in Section II, Koschmieder related bext to the dis-
tance at which a black object is just visible when viewed against the
horizon sky. The distance of visibility is given by
CMiddleton2) (4)
'ext
assuming aerosol homogeneity, uniform illumination and a 0.02 detecta-
ble contrast. Commonly it is assumed that bext = bgca , i.e., babg = 0.
Measurements of bgcat and observer visibility show good agreement with
the formula above.
Horwath and Noll conducted a study in Seattle between total
light scattering, bscat measured with an integrating nephelometer, and
prevailing visibility observed by two separate people. Their results
-------
-49-
were iii good agreement with the theoretical expression of Koschmieder
when only data for RH < 65% RH were included. Apparently the location
of the nephelometer in a heated room caused reduced RH in the light
scattering measurements. In the cases where RH < 65%, the correlation
between bscat and prevailing visibility was 0.89 and 0.91jrespectively
with a coefficient in the Koschmieder expression of 3.5 ± 0.36 and
3.2 ± 0.25, respectively. This can be compared with the theoretical
value of 3.9, indicating a slightly lower prevailing visibility than
meteorological range. Since no ideal black targets were used (only
trees, buildings, etc.)) these would have caused just such a deviation.
4
Samuels et al. conducted the most extensive tests to date of
the relationship of prevailing visibility to light scattering and
various mass concentration measures as discussed earlier.
They conclude that bsp as measured with the integrating nephelome-
ter is a good predictor of prevailing visibility and that the regression
analysis is in agreement with Koschmieder's theory. These workers noted
that there was a smaller observed prevailing visibility than that pre-
dicted from theory and bsp measurement, which they suggested was due to
non-ideal black visibility targets.
MEASUREMENTS OF SCATTERING PARAMETERS
Under support from the Environmental Protection Agency, National
Science Foundation)and the California Air Resources Board, we have
measured various aerosol scattering parameters in urban and rural loca-
tions in California, Colorado and Missouri. In all locations the in-
coming air was heated 5° to 20°C above ambient to lower relative
humidity of the sample. The measured parameters were:
bso - Scattering extinction coefficient of particles at 530 nm.
—K (Rayleigh at 530 nm = 0.15 x lO^nf1)
a - Wavelength dependence of b parameterized as
sp
b = KX"a (10)
sp
Two values of a were computed from Red-Green bsp and Blue-
Green bsp. Red is 640 nm. Blue is 430 nm. Green is 530 nm.
Scat, ratio - Ratio of half sphere back scatter to bsp from particles
at 530 nm.
The sites were:
Richmond - Northeast corner of San Francisco Bay in vicinity of
petro chemical plants.
-------
-50-
Point Reyes - Coast Guard station on cliff 150 meters above the
sea surface, 50 km NW of San Francisco.
Fresno - Central valley of California, urban agricultural site.
Hunter Liggett - Rural California site 20 km inland from ocean.
Local elevation 400 m. Local vegetation consisted of dry
grass and sparce trees.
Cal. Tec. - Site on campus in Pasadena in Los Angeles basin.
Pomona - Site at county fairgrounds in inland area of Los Angeles
b as in.
Washington Univ. - Campus site located in residential area of St.
Louis, MO.
Tyson - Rural area 25 km WSW of St. Louis.
St. Louis Univ. - Campus site in industrial St. Louis.
Henderson - Site 10 km NE of Denver.
Trout Farm - Site 8 km N of Denver.
Table 2 lists the measured values at each site. For each measure-
ment parameter, the range of that parameter containing 63% of the data
is specified. For b , the units are 10~^m and the range low to high
contains 63% of data.
bap Measurements
Using the technique described in section VI,C, measurements were
made of bap at two locations NE of Denver and three sites near St.
Louis during Fall of 1973. The measured values of the ratio of absorp-
tion to extinction are presented in Figure 5. In Denver, the absorption
to extinction ratio is very high, indicating that the aerosol heats and
stabilizes the lower atmosphere. At the three Missouri sites the
measured values are as one would expect - the rural area (Tyson) has a
less absorbing aerosol than the industrial site (St. Louis University).
Only the industrial MO site had absorption comparable to that measured
outside Denver.
The probable chemical species that produces the absorption is
graphitic carbon. Without chemical analysis for this material it is
only possible to speculate about the nature of Denver's very absorbing
aerosol. The absorption could result from:
(1) high graphitic carbon content.
(2) large concentrations of graphitic carbon particles smaller
than 0.1 urn.
-------
TABLE 2. LISTING OF MEAN AND VARIATION INCLUDING
PARAMETERS IN 11 LOCATIONS
Location b (530nm) b low b high aRG
( Units of TO'4 M"1 )
63% OF MEASUREMENTS FOR FOUR SCATTERING
aBG
Scat. Ratio
Richmond 0.4
Point Reyes 0.12
Fresno 1.0
Hunter Liggett 0.4
Cal. Tec. 1.5
1.8
Wash,' Univ. 1.58
Tyson 0.63
St. Louis Univ. 0.71
Henderson 0.31
Trout Farm 0.56
Pomona
\
0.2
0.04
0.3
0.2
0.8
0.6
Start End
Mo./Day/Hr.Yr.
u
1
1
\
1
1
1
1
1
1
1
.QTU.3
-_ .
• £• O * I •
.0+0.4
A+0 fi
. *t'\J • O
.5+0.4
.310.4
.47+0.
.80+0.
.85+0.
.65+0.
.75+0.
4
4
3
8
7
i
1
.
1
1
1
1
1
1
1
•fc_'
.7+0.
, .
—
.5+0.
-=<^^^E=i=
.3+0.
.00+0
.25+0
.2510
.15+0
.30+0
5
3
7
.4
.5
.3
.6
.5
18+5%
20+8%
16+8%
11+1%
14+_4%
14+2%
1718%
18+6%
01 1 1 \lflf.
Q/-IC/II IT)
Ol \ 'Jf 1 1 / / t.
8/29/9/72
Q/lo/q/7?
• •'/' *-/ -7 / ' *•
9/20/10/72
10/4/11/72
8/25/2/73
9/3/19/73
9/27/20/73
11/10/--/73
11/15/10/73
Of It/ 1 J/ / C
R/?S/fi/7?
O/ 1 f 1C
9/8/14/72
q/ic/in/7?
7/ 1 *J/ 1 u/ / C
10/2/8/72
10/31/15/72
8/31/9/73
9/27/12/73
10/4/15/73
11/14/--/73
11/23/11/73
I
Ul
The parameters and sites are discussed in the text. Note that in all locations the sample
air was heated 5° to 20° above ambient temperature.
-------
-52-
cr
61
Denver, Two Sites 10km
NE Of City.
11/10/73- 11/23/73
St. Louis University
9/28/73-10/4/73
Washington University
8/22/73-8/30/73
Tyson, Mo.
9/5/73-9/26/73
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
b /fb + b I
ap'^sp apJ
ure 5. Ratio of absorption to extinction by particles. Data shown
is from Fairground and Trout Farm sites NE of Denver and
three sites near St Louis Mo.
-------
-53-
(3) lack of (NH^-SO^ as a major component of Denver aerosol when
compared to that found in rural Missouri.
The ratio of absorption to filterable particulate mass can be used
to estimate an imaginary refractive index for the aerosol if a size
distribution and chemical uniformity are assumed. We believe the par-
ticles are not uniform chemically and prefer to report bs_ rather than
ri2' With this warning, the average aerosol bap at Denver was 0.35 x
10~^m~ . The imaginary refractive index, n2> given the stated assump-
tions was 0.035.
CONCLUSIONS
Comparisons can be made between our measurements at Denver and
other locations. Deliquescent salts were not detected in the aerosol
at Denver and the bsp(RH) curves were at times quite hygrophobic. The
aerosol is less water soluble in Denver than at other sites.
The aerosol had somewhat higher backscatter to bsp ratio and much
higher bap/bext than values of the same parameter at other locations.
Both measurements could be explained by a shift of the small particle
mode to smaller particles. The absorbing character of Denver aerosol
may enhance the brown or yellow color of distant white objects viewed
through the urban plume.
ACKNOWLEDGEMENTS
This research has been supported by Environmental Protection
Agency, National Science Foundation,and California Air Resources
Board funds.
REFERENCES
1. Koschmieder, H., Beitr. Phys. Freien Atm.. 12, 33-55 & 171-181
(1924).
2. Middleton, W. E., Vision Through The Atmosphere, University of
Toronto Press, Toronto, Canada (1968).
3. ACHEX Aerosol Characterization Experiment of the State of
California Air Resources Board. Prime contractor is Rockwell
International Science Center.
4. Samuels, H. J. et al., "Visibility, Light Scattering and Mass
Correlation of Particulate Matter," Report of California Air
Resources Board (1973).
5. Charlson, R. J. et al., Atm. Env. _2> 455 (1968).
-------
-54-
6. Simmons, W. A. et al., "Correlation of the Integrating Nephelome-
ter to High Volume Air Sampler," Mass. Dept. of Pub. Health (1970)
7. Thielke, et al., Aerosols and Atmospheric Chemistry. G. M. Hidy,
editor, Academic Press, New York (1972).
8. Porch, W. M., Science, 170, 315 (1970).
9. Horvath, H. , Atmospheric Environment, .5, 333 (1971).
10. Waggoner, A. P., et al., Applied Optics, 10, 957 (1971).
11. Husar, R. B., Private Communication (1974).
12. Charlson, R. J. et al., Aerosols and Atmospheric Chemistry, G. M.
Hidy editor, Academic Press, New York (1972).
13. Rasmussen, R. A. et al., PNAS 53. 1, 215 (1965).
14. Junge, C., J. Meteorology, 11, 323 (1954).
15. Charlson, R. J. et al., Science, 184, 156 (1974).
'16. Charlson, R. J. et al., Atmospheric Environment, J3, 1257 (1974).
17. Winkler, P-, Aerosol Science, 4., 373 (1973).
18. Hanel, G., Beitr. Z. Phys. Atm.. 44, 137 (1971).
19. Covert, D. S., Ph.D. Thesis, University of Washington (1974).
20. Beutell, R. G. et al., J. Sci. Inst.. 26_, 357 (1949).
21. Ahlquist, N. C., et al., J.A.P.C.A. , 17.' 467 (1967),
22. Ahlquist, N. C. et al., Patent Application (1974).
23. Ensor, D. et al., Atmos. Env.. ^, 48 (1970).
24. Quenzel, H., Atmos. Env., £ (1975).
25. Rabinoff, R. et al., J.A.M.. 12, 184 (1973),
26. Eiden, R. , Applied Optics. .5, 569 (1966).
27. Grames, G. W. et al., J.A.M., 13, 459 (1974).
28. Volz, F. E., J.G.R., 72, 1017 (1972).
29. Lindberg, J. D., Applied Optics. JJ, 1923 (1974).
-------
-55-
30. Waggoner, A. P. et al. , Applied Optics. JL2 896 (1973).
31. Bergstrom, R. W., Beitr. Z. Phys. Atm., 46, 223 (1973)
32. Lin, C. I., Applied Optics, 12, 1356 (1973).
33. Howath, H., Atmos. Environment, 3, 543 (1969).
-------
THE DICHOTOMOUS SAMPLER - A PRACTICAL APPROACH
TO AEROSOL FRACTIONATION AND COLLECTION
by
Robert K. Stevens and Thomas G. Dzubay
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
To Be Presented
At
Third Interagertcy Symposium on Air Monitoring Quality Assurance
May 18-19, 1977
Berkeley, California 94704
-------
THE DICHOTOMOUS SAMPLER - A PRACTICAL APPROACH
TO AEROSOL FRACTIONATION AND COLLECTION
Robert K. Stevens and Thomas G. Dzubay
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
INTRODUCTION
In 1972 the Environmental Protection Agency embarked on a program to
develop reliable procedures to separate and collect aerosols in two size
fractions (<3.5 microns and >3.5 microns aerodynamic mean diameter). The
size fractionation system was designed to collect the aerosols on inert
surfaces so that the mass and chemical composition could be measured with
the minimum of artifact formation.
(123) (45)
Manual ' ' and automated virtual ' dichotomous samplers were
developed and for the past three years have undergone extensive field
testing and wind tunnel studies to characterize their aerosol sampling
•. j
characteristics. Presently, these samplers are being used in 10 separate
field studies across the United States to obtain mass, sulfate, and elemental
composition data which are being compared with measurements of aerosols
collected with the Hi-Volume sampler. The results of these field and wind
tunnel tests as well as the advantages of the dichotomous sampler for char-
acterizing the aerosol content of the atmosphere are discussed in this
manuscript.
-------
DISCUSSION
Virtual Impaction Principle
Impactors that separate aerosol particles depend upon the relative
balance between inertial and aerodynamic forces. In the conventional aerosol
impactor an air'stream turns abruptly as it approaches a flat plate. Particles
with the largest inertia tend to maintain a straight trajectory and impact on
the plate while the viscous drag forces of the gas flow carry the smaller
particles along air flow streamlines. The particles collected on the plate
would ideally consist of all sizes above a well defined cut-off diameter.
However, a conventional impactor becomes a quantitative particle collector
only when the impaction surface is coated with a layer of grease to minimize
particle bounce errors ' . Unfortunately the grease can interfere with
some of the required chemical analyses and may become completely covered with
particles causing particles to no longer be quantitatively collected.
In the virtual impaction collection method, instead of having the larger
particles collect on grease coated plates, the larger particles are impacted
into a slowly pumped void and collected on a filter. Since a small constant
fraction of the total flow is being pulled through the void, the portion of
fine particles contained in the large particle fraction is directly related
to the ratio of the inlet flow and the flow rate in the large particle
collector. To reduce the small particle contamination, two stages of virtual
impaction are operated in series as shown in Figure 1.
During the initial phases of development of the virtual impactor, it
was decided that dividing and collecting aerosol into two size ranges (hence
the name dichotomous sampler) would provide the optimum amount of information to
distinguish primary from secondary pollution sources. Separating the particles
- 2 -
-------
into more than two fractions increases the complexity of the sampler, reduces
the amount of aerosol per stage and increases the opportunity for losses of
particles in the sampling train.
The main advantages of the virtual impactor over conventional aerosol
size fractionation designs are:
1. Particle bounce and reetrainment problems associated with
ungreased impactors and the chemical interferences caused by
the required grease coating are eliminated.
2. Sufficient quantities of aerosol particles are collected on
inert surfaces that are ideally suited for subsequent gravi-
metric, elemental, and chemical analysis.
3. Aerosols are uniformly deposited on filters making an ideal
sample for X-ray fluorescence (XRF) elemental analyses and
beta-gauge mass measurements.
Collection Surfaces
The filter medium of choice is high porosity Teflon with 1 ym pores
for the following reasons:
1. Collection efficiency for particles above 0.01 ym is greater
than 99%(8).
2. Extremely stable mass for high gravimetric accuracy.
3. Negligible tendency to absorb or react with gases.
4. Minimal impurities to interfere with analyses for chemical and
elemental species.
5. Low mass per unit area. (Desirable for gravimetric, XRF, and
(5-gauge measurements).
- 3 -
-------
In addition Teflon filters are ideal for the collection of sulfuric
acid since there is little if any interaction with the filter medium. In
laboratory experiments when between 1.0 and 100 ug of 0.3 ym droplets of
sulfuric acid were deposited on Teflon filters, recoveries of >90% of the
sulfate and equivalent acidity have been demonstrated.
Gravimetric determination of the aerosol mass collected on Teflon is
significantly less affected by changes in relative humidity than for glass
fiber, quartz, and esters of cellulose types of collection surfaces.
For aerosol sampling, the Teflon filters must be supported or mounted
in a manner to provide adequate structual strength. In extensive tests
with the manual dichotomous sampler, we have successfully used 37 mm
diameter Fluoropore (Millipore Corp. ) filters, which consist of a Teflon
membrane bonded to a polyethylene net for support. Recently, a Teflon mem-
brane filter supported with a thin annular polyester ring has become avail-
able from Ghia Corp, Pleasonton, CA, and is currently undergoing field
testing. Such a filter is expected to have superior flow and loading
characteristics and because of its lower mass per unit area, it is expected
to be superior for beta gauge mass determinations and XRF analyses.
Inlet Design
Conventional aerosol samplers use a variety of aerosol intake designs
ranging from the gabled roof used in the Hi-Volume sampler to circular "hat"
designs. With these designs, the intake sampling efficiency is variable
9 10
with wind speed and direction. Davies ' has investigated the effects that
Mention of commercial products does not constitute endorsement by the
U. S. Environmental Protection Agency
-------
particle inertia and gravitational settling have upon the performance of
aerosol inlets and found that quantitative sampling becomes increasingly
more difficult to achieve as the particle size increases above 10 - 20 ym.
Therefore, we chose to design an inlet which rejects particles above 20 pm.
Although the Hi-Volume sampler does collect particles larger than 20 jjm,
the efficiency for collecting such particles is highly dependent upon the
wind speed and wind direction relative to the orientation of its gabled roof.
Under EPA Grant 804190 Andrew McFarland at Texas A&M University has
developed an inlet, shown in Figure 2, that is compatible with the dicho-
tomous sampler. McFarland has shown through wind tunnel tests that the
particle sampling efficiency of this inlet remains relatively constant
at wind speeds between 4-8 km/hr. Figure 3 is a graph showing inlet
particle penetration as a function of particle size at a wind speed of
4 km/hr. Wind tunnel tests are being continued at Texas A&M to further
characterize the aerosol sampling efficiency of this new inlet.
Flow Control
For quantitative aerosol sampling, it is necessary to maintain a constant
flow rate through the dichotomous sampler and filters. The required flow
rate controller must compensate for changes in the flow resistance that
occur as particles accumulate on the filters during the sampling interval.
Flow rate controllers are available that operate on the following principles:
1. anemometer sensor and variable power to pump,
2. pressure differential flow sensor and motor controlled needle
valve between filter and pump, and
3. use of a differential flow controller on the pump exhaust.
- 5 -
-------
Of the above approaches, the use of the differential flow controller
on the pump exhaust appears to be the most promising for cost effective,
routine field use. A schematic of this type flow control system is shown in
Figure 4. Here a differential pressure regulator in the exhaust line from
the pump is used to maintain a constant pressure differential across a fixed
orifice thereby maintaining a constant flow rate through the system. Because
only a few inexpensive but highly reliable components are needed, this approach
provides the needed ruggedness, reliability and low cost needed for an aerosol
sampler designed for routine field use. The differential flow control system
has been tested by McFarland between -20 to +40° C, and less than a 5% change
in flow was observed over the temperature range at a flow rate of 14.0 liters/
minute. Side by side comparisons of the differential flow control system
with a commercial anemometer flow controller incorporated into two identical
dichotomous samplers were performed at Research Triangle Park, N. C. Both
controllers maintained flow constant to better than 5% for pressure drops
across the filter up to 25 cm of Hg.
Stacked Filter Sampler
Stevens and Pzubay have previously described an alternative type of
dichotomous aerosol fractionating sampler that consists of two filters in
series. The first is a Nuclepore filter with 12 urn pore diameters and the
second is a Teflon filter with 1 urn pores. The 12 \im Nuclepore filter has
been characterized by Parker et al. to shoxj that when operated at the
appropriate flow rate, the collection efficiency curve for particles with a
specific gravity of 2 approximates the ACGIH criteria for respirable sampling
and has a 50% cut point at an aerodynamic diameter of 3.5 ym . However
6 -
-------
as the specific gravity varies, there is a slight variation in the collection
efficiency curve vs aerodynamic diameter. Tests in ambient air indicate
that the fractionation curve is not affected by particle loading
With this simple combination of two filters in series, one has a low
cost, lightweight aerosol sampler that can separately collect the respirable
and nonrespirable particles. By adding additional chemically treated stages
(12)
as shown in Figure 5, one can collect H_S and S07 . Such a sampler can
simultaneously collect both gasses and particles to enable one to determine
the relationship between gaseous and particulate sulfur.
Typical Results
Dichotomous samplers of the virtual impaction design have been operated
in a number of geographical areas, and extensive sampling has been conducted
in St. Louis, Missouri, and Charleston, West Virginia. Figure 6 shows a
comparison of the total (fine plus coarse) mass concentrations determined
from Hi-Volume and dichotomous samplers operated in St. Louis during the
summer of 1975. In determining the mass values, the Fluoropore filters used
in the dichotomous samplers were weighed to a precision of 10 ug using an
electrobalance in a room with the relative humidity adjusted to 40%. Immedi-
210
ately before the filters were weighed, they were passed in front of a Po
radioactive source to remove any electrostatic charge. The glass fiber filters
used in the Hi-Volume samplers were weighed using a mechanical balance. The
results of this study are shown in Figure 6 and indicate the potential of the
dichotomous sampler to be used as a substitute for the Hi-Volume sampler for
determining mass concentrations. However, in dusty locations when turbulent
winds suspend significant amounts of particles larger than 20 ym in the
-------
atmosphere, one would not expect to find such good agreement between samplers
since the Hi-Volume sampler collects significantly more of the larger parti-
cles.
In addition to being suitable for gravimetric analyses, the Teflon
collection surfaces of the dichotomous sampler are well suited for making
elemental and chemical analyses. Using an x-ray fluorescence spectrometer,
the samples can be analyzed nondestructively for elements with elements with
atomic numbers above 13 ' ' ' . With the Dionex Ion Chromatograph ,
sample extracts can be analyzed for a wide variety of ionic species including
sulfate, sulfite, and nitrate. Using the Brosset thorin titration, the Gran
= +
titration and an ion selective electrode, one can determine SO, , H , and
NH, concentrations . Such analyses are important in determining sulfuric
acid, ammonium sulfate and ammonium bisulfate in the sample.
Table 1 shows the average concentrations deduced from twenty sampling
periods in St. Louis using a dichotomous sampler and an x-ray fluorescence
analyzer. One notes that in the fine particle range, sulfur is the pre-
dominant species. Analysis of the samples using ESCA (photoelectron
spectroscopy) reveals that the sulfur is in the form of sulfate. The paucity
of metals in the sample, especially at the rural site, indicates that metal
sulfate compounds are of minor importance. The measurements of ammonium and
hydrogen ions indicate that the sulfate is usually in the form of (NH.)„ SO..
Table 1 also shows that the sulfur in St. Louis occurs predominantly in
the fine particle fraction. Figures 7 and 8 which are plots of fine and
coarse particle sulfur and mass fractions collected in Charleston, West
Virginia, also show that the sulfur occurs predominantly in the fine particle
fraction. This same pattern has also been observed for samples collected in
-------
Table 1. Mass and Percentage Composition of Size Fractionated
St. Louis Aerosol Averaged over the Period
from August 18 to September 7, 1975
Si
S
K
Ca
Ti
Fe
Zn
Br
Pb
Urban
Fine (%)
29 yg/m3
1
12.5
0.4
0.7
1.1
1.4
0.35
0.33
2.2
Coarse (%)
22 yg/m3
8
1.4
1.2
8.2
2.0
4.8
0.20
0.16
0.60
Fine (%)
3
26 yg/m
0.5
12.6
0.3
0.5
<0.1
0.3
0.13
0.06
0.51
Rural
Coarse (%)
15 yg/m3
4
0.9
0.9
4.2
0.2
1.3
0.15
0.04
0.11
located at the Missouri Botanical Garden in St. Louis.
Located in an agricultural area in Illinois, 40 km south of St. Louis.
- 9 -
-------
Los Angeles, Denver, New York, Philadelphia, Milford, Mich., and Durham, N.C.
The only exception to this pattern was observed when a sampler was operated
very near a fertilizer plant in East St. Louis, Illinois. There the sulfur
in the coarse particle fraction sometimes exceeded that in the fine fraction.
A high calcium content in the sample indicated that the particles consisted of
CaSO,, and this was confirmed by microscopic examination.
Figures 9 and 10 show a comparison between XRF measurements for sulfur
(expressed as sulfate) and measurements for sulfate on the same set of fine
particles collected in Charleston, W. V. using the ion chromatograph and
the thorin titration. The solid lines illustrate the case of perfect agree-
ment, and the dashed line represents the results of linear regression
analyses. The closeness of the two lines in both Figures 9 and 10 illustrates
the excellent agreement between methods. For example in Figure 9 the
expression describing the linear regression is y = 0.983 x -323.1 and for
Figure 10 the expression is y = 0.962 x +294.5. The units for x and y are
nano-equivalents of sulfate. The correlation coefficients for figures 9 and
10 are 0.981 and 0.989 respectively. This comparison of the sulfur concentration
by XRF with sulfate measurements with the ion chromatograph strongly suggests
that 85% or more of the sulfur in particles less than 3.5 microns in aero-
dynamic diameter, is in the form of sulfate. Since the ammonia concentration
in the same samples also was found to be present in the same equivalent con-
centrations as the sulfate, we conclude the sulfate is in the form of
ammonium sulfate. The chemical form of sulfur in the fine particle fraction
in Charleston, W. V- appears (see above) to be the same as observed in aerosols
collected and analyzed in St. Louis, MO.
Ten automated dichotomous samplers (ADS) located in the St. Louis, MO
area and part of EPA's Regional Air Monitoring System^ ' (RAMS) have been
- 10 -
-------
in continuous operation since March of 1974. At these same sites from March
1975 to March 1977, Hi-Vols were operated approximately every 3rd day. The
fine and coarse fraction of aerosols collected with the ADS were analyzed
for mass and elemental composition (by x-ray fluorescence) while aerosols
collected with the Hi-Vol were analyzed for mass and sulfate content
(methyl-thymol blue method for sulfate).
Figure 11 presents time series plots for September through December
1975 for two RAMS's stations, 106 (located in Busch Botanical Gardens) and
122 (located 45 km North of 106) of total sulfate concentrations deter-
mined for aerosols collected by the Hi-Vol and dichotomous samplers.
Although dichotomous sampler data are available on a daily basis, only those
points are plotted which correspond to days for which Hi-Vol data were taken.
The data are typical of that obtained in St. Louis during this period.
Inspection of the plots reveals three salient qualitative features of
the data:
1. The Hi-Vol sulfate data are characterized by spikes in concen-
tration which are local and show no apparent correspondence from
station to station.
2. The Hi-Vol sulfate data show concentrations which are on the
average higher than those obtained with the dichotomous sampler.
Preliminary analysis indicates that this conclusion still holds
even if outliers during spike periods are not included. This
observation is made more apparent by inspection of Figure 12
in which plots for the ratios of Hi-Vol sulfate to total dicho-
tomous sulfate are presented for the same two stations. The
average ratio, 1.25, at station 106 is lower than that of 1.44
obtained at station 122.
- 11 -
-------
3. The data at a single station for both methods are fairly well
correlated.
Figure 13 shox^s the dichotomous sampler data in more detail in that
simultaneous time series plots for the sulfate found in both fine and
coarse fractions are presented. It is clear from these data that in the
particle size range below 20 microns the overwhelming proportion of sulfate
is contained in the fine fraction (below 3.5 microns).
A brief summary of St. Louis sulfate concentrations is shown in
Table II for eight RAMS stations monitoring sulfate by both the Ei-Vol and
dichotomous sampler methods. For a composite of 8 stations the time averaged
Hi-Vol sulfate concentration was 10.3 pg/m while that for the dichotomous
3
sampler (fine plus coarse fractions) was 7.8 yg/m . For every station the
time averaged concentration measured by the Hi-Vol exceeded that measured .by
the dichotomous sampler, indicating a discrepancy betw.een the two methods
ranging from 23 to 53 percent (less if outliers are excluded). No firm
explanation of this preliminary result is offered. It is suggestive, how-
ever, of either aerosol sampling efficiency variations or possible sulfate
artifact formation on the glass fiber filters of the Hi-Vol. Note that if
the measured difference were real, there would seem to be a significant
sulfate contribution from particles of greater than 20 micron aerodynamic
size. (Such particles might be unimportant in so far as human health hazards
are concerned.) The dichotomous sampler data show that on 94% of all the
days sampled, the percentage of sulfur in the fine fraction (below 3.5
microns) was no less than 70%.
Table II also shows that the time-averaged Hi-Vol data correlate well
with the dichotomous sampler data. A calculation of the Spearman rank-
- 12 -
-------
order correlation coefficient yields the value ..83 at a significance level
of >.99.
TABLE II
SUMMARY OF ST. LOUIS SULFATE CONCENTRATIONS (ug/m )
DETERMINED BY TWO METHODS FOR THE PERIOD SEPT THRU DEC 1975
RAMS Station Avg Avg Total Ratio Hi-Vol % Days When
Number Hi-VOL Dichotomous Total Dichotomous Fine S/Total S>0.?0
*106
108
112
115
118
120
*122
124
11.0
12.7
10.6
9.6
9.9
9.9
8.9
9.8
8.9
10.0
9.2
7.8
6.6
7.3
6.2
6.4
1.25
1.27
1.23
1.23
1.50
1.35
1.44
1.53
97
79
95
99
83
100
94
99
COMPOSITE
EIGHT STATIONS 10.3 7.8 1.32 94
SPEARMAN RANK-ORDER CORRELATION COEFFICIENT = .83
P<.01
*STATIONS PRESENTED IN TIME SERIES PLOTS
Figure 14 shows (corresponding) time series plots for total mass
collected by Hi-Vol and dichotomous samplers. Many of the same qualitative
features noted above for the sulfate measurements are also apparent in the
mass data: spikes in the Hi-Vol measurement and a higher, (in some cases
much higher) value obtained by the Hi-Vol. The time-averaged value for the
ratio of Hi-Vol mass to total dichotomous mass at station 106 was 1.8. The
corresponding ratio at station 122 was 1.2. It should be recalled that
the sulfate ratios were higher at station 122 than at station 106. Further
work is needed to clarify these relationships for both sulfate and mass.
- 13 -
-------
Manual and Automated Designs: Second Generation
The prototype manual and automated dichotomous samplers have been
throughly tested, and their reliability as aerosol fractionaters is well
proven. However, the prototypes were expensive to fabricate due to the
machining precision required. In addition, the prototypes required an
expensive flow controller. The flow control problem has been solved by
employing a low cost differential flow controller, but the fabrication and
assembly of the critical flow components is still a problem. EPA has
recently funded Beckman Instruments, Fullerton, CA, to design and fabricate
a manual dichotomous sampler that could be obtained for less than $1,900 and
an automated sampler that could be obtained for less than $4,500. For certain
components Beckman proposes to use metallized molded plastics where the
components are made in two halves which are then adhesive-bonded together.
Since concentricity of inlet nozzles and receptor orifices in the impactor
stages are critical, each impactor stage is made of one piece. Then the
assembly of parts becomes non-critical. Nozzles and receptors were fabri-
cated as single units that can be easily removed for cleaning without affecting
allignment. The manual sampler built by Beckman will be designed for direct
installation in exposed locations with no additional housing required. The
instrument will be housed in a seamless glass reenforced plastic enclosure
which will have an access door for convenient changing of the filters or
cleaning the virtual impactor. In addition to the work of Beckman, Sierra
Corporation, Carmel Valley, CA, is currently selling a dichotomous sampler
developed by ERC, St. Paul, Minn., under EPA Contract 68-02-1744. Sierra
couples the ERC sampler with their flow controller and special inlet system
and markets the complete aerosol collection device at a cost of about $3500.
-------
The automated system is similar to that of the manual sampler. Common
components include inlets, virtual impactors, flow controllers and air trans-
port and external housing. The sampler changing mechanism will be adapted
from the sample changer mechanism developed for EPA by Lawrence Berkeley
(4 5)
Laboratory ' . Magazines containing up to 36 pairs of filters will be
installed through an easy-access service port. The changing of the filters
will proceed according to instructions from the control module. The control
module will be composed of a microprocessor and pressure drop sensing element.
When the pressure across the filter exceeds a pre set value, the filters will
automatically be changed. In most cases the filters will be changed before
the pressure drop exceeds the pre-set level, ie. once every 24 hours. How-
ever during period of air stagnation, particulate levels may exceed several
hundred micrograms per cubic meter. In these instances when the filters begin
to clog, the filters will be changed before the flow decreases by more than
a few percent of the initial setting.
This new family of aerosol samplers is a marked departure from previous
approaches of aerosol sampling. Years of prototype testing, evaluation and
intercomparisons have been performed. These tests have culminated in the
current design concepts soon to be implemented by several instrument manu-
facturers and should provide the various air pollution control agencies with
a powerful tool to aid in understanding sources and transport of atmospheric
aerosols.
-------
ACKNOWLEDGMENTS
The authors are indebted to Dr. Andrew McFarland of Texas A&M University
for providing the data and drawings shown in Figures 2 and 3. Also the
authors acknowledge the assistance of Dr. George Russwurm and Dr. Dwight
Rickle of Northrop Services Inc. for providing some of the chemical analyses
cited in this report. In addition we thank the West Virginia Air Pollution
Control Board for operating the manual dichotomous sampler from which some
of the data in this report was derived. We acknowledge Dr. S. Friedman and
Joseph Nader, System Science, Inc., Chapel Hill, N. C. for assistance in
preparation of Figures 11-14 and interpretation of data related to measure-
ments made in St. Louis, MO. The x-ray fluorescence analyses of samples
collected in St. Louis with the automated dichotomous samplers were perfromed
by Lawrence Berkeley Laboratory.
- 16 -
-------
REFERENCES
1. Stevens, R. K. and T. G. Dzubay. "Recent Development in Air Particulate
Monitoring," IEEE Trans, on Nucl. Sci. NS-22. 849-855 (1975).
2. Dzubay, T. G. and R. K. Stevens. "Ambient Air Analysis with Dichotomous
Sampler and X-Ray Fluorescence Spectrometer," Environ. Sci. Technol. j?,
663-668 (1975).
3. Dzubay, T. G. and R. K. Stevens, "Application of the Dichotomous Sampler
to the Characterization of Ambient Aerosols," in X-ray Fluorescence
Analysis of Environmental Samples (Ann Arbor, Michigan:Ann Arbor Science,
1977) 95-105.
4. Goulding, F. S., J. M. Jaklevic and B. W. Loo. "Fabrication of Moni-
toring System for Determining Mass and Composition of Aerosols as a
Function of Time," Environmental Protection Agency, Report No. EPA-
650/2-75-048 (1975).
5. Loo, B. W., J. M. Jaklevic and F. S. Goulding. "Dichotomous Virtual
Impactors for Large Scale Monitoring of Airborne Particulate Matter,"
in Fine Particles, B. Y. H. Liu, Ed. (New York: Academic Press, 1976),
pp. 311-350.
6. Wesolowski, J. J., W. John, W. Devor, T. A. Cahill, P. J. Feeney,
G. Wolfe and R. Flocchini. "Collection Surfaces of Cascade Impactors,"
in X-Ray Fluorescence Analysis of Environmental Samples, T. G. Dzubay,
Ed. (Ann Arbor:Ann Arbor Science, (1977) 121-131.
7. Dzubay, T. G., L. E, Hines and R. K. Stevens. "Particle Bounce Errors
in Cascade Impactors," Atmos. Environ. 10, 229-234 (1976).
8. Liu, B. Y. H. and G. A. Kuhlmey. "Efficiencies of Air Sampling Media"
in X-Ray Fluorescence Analysis of Environmental Samples (Ann Arbor:
Ann Arbor Science, 1977) 107-119.
9. Davies, C. N., Royal Soc. London Proc., Sec. A., Vol. 279, No. 1378:
413 (June, 1964).
10. Davies, C. N., Brit. J. Appl. Phys., Sec. 2, Vol. 1 (1968).
11. Parker, R. D., G. H. Buzzard, T. G. Dzubay, and J. P. Bell, "A Two
Stage Aerosol Sampler Using Nuclepore Filters in Series," Atmospheric
Environment (1977) to appear in May or June.
12. Lorenzen, J. A., "Environmental Monitoring Device for X-Ray Determination
of Atmospheric Chlorine, Reactive Sulfur, and Sulfur Dioxide," in Advances
in X-Ray Analysis, Vol. 18 (New York:Plenum Press., 1975) 568-578.
- 17 -
-------
REFERENCES cont'd
13. Jaklevic, J. M., F. S. Goulding, B. V. Jarrett and J. D. Meng.
"Applications of X-Ray Fluorescence Techniques to Measure Elemental
Composition of Particles in the Atmosphere," in Analytical Methods
Applied to Air Pollution Measurement. R. K. Stevens and W. F. Herger,
Eds. (Ann Arbor, Michigan:Ann Arbor Science Publishers, 1974), pp. 123-146,
14. Mulik, J., R. Puckett, D. Williams, and E. Sawicki, "Ion Chromatographic
Analysis of Sulfate and Nitrate in Ambient Aerosols," Analytical
Letters 9. 653-663 (1976).
15. Brosset, and M. Fern, "Man Made Airborne Acidity and Its Determination,"
(IVL Publications B-314A. Gothenburg, Sweden 1976).
16. Novakov, T., S. G. Chang, R. L. Dod, H. Rosen, "Chemical Characterization
of Aerosol Species Produced in Hetrogeneous Gas-Particle Reactions",
Lawrence Berkeley Laboratory Report No. LBL-5215 (1976).
17. Myers, R. Lee and James A. Reagan, "The Regional Air .Monitoring System",
St. Louis, MO, Proc. International Conference on Environmental Sensing
and Assessment, Los Vegas, Nevada, Sept. 14-19, 1975.
18. See for example N. H. Nie, C. H. Hull, J. G. Jenkins, K. Steinbrenner,
D. H. Bent, Statistical Package For the Social Sciences (New York
HI 1 1 . 197Si.
- 18 -
-------
141/min
FROM INLET
VIRTUAL
IMPACTOR
NO. 1
VIRTUAL
IMPACTOR N0.2
COARSE
PARTICLE
FILTER
FINE
PARTICLE
FILTER
TO
SERVO
PUMP
13.6 l/min
AND
PUMP
0.4 l/min
&&S&S$s8S5&&isQQ&S!S&&!$SG9Sl5Gimf5&a
SCALE, in
Figure 1. Cross sectional view of virtual impactor hav-
ing two substages.
- 19 -
-------
SCALE, in SCALE, cm
OTIT 0
-8
FLOW
BUG SCREEN
16X16 MESH
SHIELD
FLOW TO DICHOTOMOUS SAMPLER
Figure 2. Aerosol inlet for dichotomous sampler.
y -20-
-------
100
s
g 80
60
UJ
w 40
s: 20
s
CO
3 5 7 10 20 30
AERODYNAMIC PARTICLE DIAMETER, pm
Figure 3. Sampling efficiency for" the aerosol inlet shown in Figure 2.
AMBIENT AEROSOL FLOW
0.4 LPM
VENT TO
ATMOSPHERIC
PRESSURE
TWO-STAGE
VIRTUALIMPACTOR
FRACTIONATING INLET
BALLAST
VOLUME
FINE
ROTAMETER
nuiMivicitn
BALLAST
VOLUME
BALLAST
VOLUMES
NEEDLE
VALVE
ATMOSPHERIC
PRESSURE
REFERENCE
117 VAC _Jr
1 1
ELAPSED TIME
INDICATOR
1
ON-OFF
TIMER
PRESSURE
SWITCH
WITH
117 VAC MANUAL
POWER RESET
CONSTANT
DIFFERENTIAL
PRESSURE
FLOW
CONTROLLER
^CONTROL
VALVE
AP = 3 PSI
@ 13.6 LPM
FINE PARTICLE
ROTAMETER
Figure 4. Schematic view dichotomous sampler with constant flow rate system
which uses a pressure regulator to maintain a constant pressure differential
across a fixed orifice.
- 21 -
-------
ro
ro
IT
L-LJJ
COARSE
PARTICLES
(12/umNP) |
FINE PARTICLES
(0.2AimNP)
H2S(AgN030N
CELLULOSE) \
S02(NaOH ON
CELLULOSE) ^
100
50
TO PUMP
6.5liters/min
Figure 5. Schematic view of stacked filter aerosol and gas sampler.
C = 0.94
Y • 0.977 x-1.7'
I I I I I
I' I I
0 50
FINE +COARSE MASS,
Figure 6. A comparison of total mass concentrations deter-
mines using high volume and dichotomous samplers operat-
ed at the St. Louis Botanical Gardens between August 18
and September 7, 1975.
100
-------
C.U
74
fc*t
22
20
18
„ 16
E
"s 14
£ 12
<
jj; 10
5 8
6
4
2
4-
o
1
I
~ FINE FILTERS
— COARSE FILTERS
—
—
—
—
—
- —
—
-
l
;"l
1 1
LL_
i
1 1
i _L
-
1
i ' . '
!i :
!
.
1
i
1
CC CC > > >• S Z Z
2=S:<<<=3:3=> =
< en e
•- c
" rr-
«J CC
Figure 7. Fine and coarse particle sulfur measurements for aerosols collected
in Charleston, WV during 1976 using a dichotomous sampler.
ibu
150
140
130
120
4 110
o 90
5 80
o 70
^ 60
H 50
o
H 40
30
20
10
i i
— i i
FINE FILTERS
_^ COARSE FILTERS
—
—
_
—
w
r
IT
r l
i
1 if J 1
& ft
U7 H 1
ffi fr *
JK
L b
ii
H
in
i
i
f t&t
ifrt
1 IK
H
1 Ki I
1
1
|
\
•<
\
I
<
1
|
f
i 1
i
,'
li
:
I
1
!
1
L i
P
I i
CC CC >- >• > Z "Z. 2
:i i
' '
i
1 1
i L_
1
1 i
h
I
,
'
1
3
j — i
.
^__^
—
_—
—
'__
—
1
1
1
1
—
—
_
—
•"•"•
J C2
a- a. «j «j < :a => r> =3 =3 :=3 :=]
ea o ^C3o>cnenenencnoc
CM e*> ~ ^ CO «- CM T-CM
DATE
Figure 8. Fine and coarse particle mass measurements for aerosols collected
during 1976.
-------
CO
a
x 6
5 5
V)
a
LU
CD
2
cc
I
DICHOTOMOUS FILTERS- FINE
CHARLESTON,^/. VA.
APRIL TO AUGUST, 1976
01234567
EDXRF (NANO-EQSULFATE) X 10-3
Figure 9. A comparison of fine particle sulfur analyses using an ion exchange
liquid chromatograph for sulfate and an x-ray fluorescence spectrometer.
x
rx 5
CO
o
LU
cc
x
a
I I I I
DICHOTOMOUS FILTERS- FINE
CHARLESTON, W. VA.
APRIL TO AUGUST, 1976
01 23456
BROSSET THORIIM (NANO-EQ SULFATE) X 10'3
Figure 10. A comparison of fine particle sulfur analyses using an x-ray
fluorescence and thorin titration methods.
- 21+ -
-------
I
ro
S
a
S°
in o.
S04: HI-VOLUME SflMPLES VERSUS DICHOTOMOUS SRMPLES FOR STRTION 106
HI-VOL JNSTRUHENT
DICNOTOHOUS SOHPLER
^48.00 257.00 268.00 275-00 284.00 293-00 302.00 311.00
1975 JULIRN DRY
320.00 329.00 338.00 347.00
356
00 365.00
§
o
IA
•2l
o
•w o
o •
V3 O
S04s HI-VOLUME SflMPLES VERSUS DICHOTOMOUS SRMPLES FOR STflTION 122
HI-VOL INSTRUflENT
DICHOTOMOUS SflMPLER
°248.00 Zfe7.00 266.00 275.00 284.00 293-00 302.00 311.00 320.00 329.00 338-00 347.00
1975 JULIRN OflY
§
365.00°
u
i
.o
JO »
o
-------
(X)
ON
O
5
OC
SULFRTE RRTIO: HI-VOL OVER DICHOTOMOUS 106
O «•
SULFHTE RfiTIO: HI-VOL OVER DICHOTOMOUS 122
^43-00 257.00 266.00 275.00 284.00 293.00 302-00 311-00 320.00 329.00 338-00 347.00 356-00 365-00°
1975 JULIRN DRY
^8.00257-00266-00275-00284.00293-00302-00311.00320.00329-00339-00347.00356.00365-00°
1975 JULIRN DRY
Figure 12. Ratio of Hi-vol sulfate to total dichotomous sulfate at two St Louis sites.
-------
IX)
3=
o °
-1 «o"
z
z
o
""" O
ff O
Jf «o"
z
Ul
u
z
o
o
a: °
to
UJ
is
> «>~
$
er
Q O
SULFURs DRILY
FINE FRflCTION
CORRSE FRflCTION
?\ **
\
/ : i
' ° <
1 F V
i / a
1 j X
y \
**\y ^^r^
or"
flVERflGES FOR STflTION 106
d NORHHL OflTH « BELOH DETECT LIMIT
X MISSING DflTfl X MISSING * UNDETECTED
O
I l\ _,
M 1 \ Jp
1 i' i /* i * / 1
\ * • /\'ii
V j\ l\^ A AM /^ \
i rf '« nr^ ^ '/ 'i / 1 / \ •' 1 / '• '' \ '*
i •' 1 r j; V W m-» ' \ ' '' / • '
V^^Si^* "^r \_~* ^"w^v '3-8^^-5i.*HwJi;^l«^ V^V"1 »I**^lli_JS^»e*. ^••e(''l!' 7^1 An
o x*"^*tr o"^ B ^fi^^w^ ^ST-8^13130^ ^S^'^VB 6 1B»«*»*»a-B ^*H«*V
0
"oo
o
o
-------
ro
CD
o
UJ.
J HI-VOLUME SRMPLES VERSUS DICHOTOMOUS SfiMPLES FOR STRTION 106
HI-VOL INSTRUHENT
DICHOTOMOUO
vl O
O
ir •
c o
cr>
r o
o
MQSS: HI-VOLUME SRMPLES VERSUS DICHOTOMOUS SflMPLES FOR STRTION 122
-•-• Hi-VOL INSTRUMENT
— 01CHOTOMOUS SflUPUER
IB-
3*.
o
o '
Z57.0C 266.00 275.00 284.00 293.00 302.00 311.00 320.00 329-00 33'3.0C 347.00 355.00
1975 JULIflN DRY
257.00 266.00 275.00 284.00 293.00 302.00 311-00
1975 JULIAN DRY
320.00329.00338-00
347.00356.00
o
§1
Lei?
i
! O
-I?
:n O
o
O
3*
:a
O
O
O
§1
o
o
365.00
Figure 14. Comparison of total mass concentrations measured by Hi-vol and dichotomous sampler
at two St. Louis sites.
-------
STRATEGIES FOR IMPROVED SUSPENDED PARTICIPATE MONITORING*
FINAL REPORT
Prepared by
The Standing Air Monitoring Work Group
January 1977
for presentation at the Third Interagency Symposium on Air Monitoring
Quality Assurance, Marriott Inn, Berkeley, California, May 18 and 19, 1977
-------
INTRODUCTION
The attainment and maintenance of the national ambient air quality
standards (NAAQS) for particulate matter is affected by numerous
factors. They include the degree of control of the sources that emit
suspended particulates, particulate transformations and meteorological
»
variables.
Sources of total suspended particulates (TSP) may be categorized
in three major groupings: (1) traditionally controlled sources of
particulates, (2) non-traditionally controlled sources, and (3) trans-
ported natural and man-made particulates.
Traditionally controlled sources include industrial process
emissions (both stack and fugitive emissions), fuel combustion sources
and solid waste disposal activities. Traditional sources of particu-
late emissions have significantly reduced their emissions over recent
years, but still contribute significantly to ambient levels in heavily
industrialized communities. Fugitive emissions from the majority of
traditional sources also continue to be a significant problem.
Included in the category of non-traditionally controlled sources
are fugitive dusts from construction and demolition operations, auto
exhaust particulate emissions, road dust re-entrainment, dust from
unpaved roads and particulates from other miscellaneous community
activities. Non-traditionally controlled sources have been found to
contribute TSP emissions to such extent that they have been primarily
responsible for the failure to achieve NAAQS in many areas.
-------
2
Transported participates, both natural and man-made, have been
regarded, in certain areas, as the prime contributors to background
TSP levels. The term background is used here to denote ambient parti -
culate levels measured in non-urban areas. These background particu-
late levels can be man-made, as well as of natural origin, and are
important to know in order to properly design air pollution control
plans. Usually the background levels measured at these non-urban area
sites can be attributed to man-made particulates from distant urban
sources. In the midwest and west the effects are considered to be
minimal. Recognition, however, must also be given to particulate pol-
lution levels resulting from natural processes. The natural pollutants
include both primary and secondary particulates and are generally con-
sidered uncontrollable. A primary particulate is emitted directly to
the atmosphere and exists in the form in which it was emitted. Examples
are sea salt, entrained soil, and aeroallergens. A secondary particu-
late is formed in the atmosphere by various forms of reactions. Included
as examples of natural secondary particulates are sulfates from sulfur
oxides and ammonium salts from ammonia.
Transported man-made pari'iculates likewise can be classified as
primary and secondary particulates. Studies of the transport distance
of particulates indicate both small- and large-scale transport with
travel distances ranging from several to 100 kilometers. For effective
air resource management in areas with high particulate concentrations
attributable to transported man-made particulates, large area control
strategies might be necessary to minimize the effect of these trans-
ported particulates.
-------
3
The Federal Reference Method (FRM) for measuring suspended parti -
culate matter specifies use of the high-volume sampler. The sampler
collects greater than 99.9 percent of all the particulates 0.3 ym or
greater in diameter but does not separate the particulates collected
into various size fractions. Fine particles (roughly defined as less
than 2 vim) have been a topic of concern for many years. They are
important for several reasons: they remain suspended for long time
periods, limit visibility by scattering more light per unit mass than
larger particles, and account for one-third to one-half of the total
mass of the collected suspended particulates. Also, fine particles
are important because they comprise most of the TSP as far as number
of particles and surface area on which toxic compounds can be attached
or formed. Fine particles penetrate deeply into the respiratory system,
creating a probable health hazard because of their own toxicity and
because of any attached toxic substances.
In addition to particle size, consideration has recently been
given to the determination of the chemical composition of particulates.
Recent preliminary data from EPA studies have shown that specific fine
particulates species such as sulfates and nitrates produce more of an
adverse health effect than total fine particulates. However, analysis
of the fine particulate fraction of the total suspended particulate
sample indicates that the chemical composition varies from one area of
the country to another so that fine particle monitoring strategy or fine
particle control alone is not the solution. The approach being taken by
-------
4
EPA places greater emphasis on the study of selected categories of
fine participate matter such as lead, sulfates, and organics with
continued reliance on the current particulate matter NAAQS for over-
all control of particulates.
Currently, a national ambient air quality standard for lead is
being developed. A fraction of the exposed high-volume sampler filter
will be used for lead analysis. In the future national air quality
standards may be developed for other toxic particulate fractions. With
the exception of the lead standard, however, it is not anticipated that
any other standard that considers particulates will be promulgated with-
in the next five years. In addition, the current national standard for
particulate matter is not expected to be rescinded or replaced in the
next several years.
No practical procedures for determining equivalency of alternative
methods to the particulate FRM currently exist or are anticipated in
the near future. As a result, the high-volume sampler is the only
currently acceptable method.
Considering the above, SAMWG believes the future monitoring strat-
egy for particulate matter should include the continuation of the use
of the high-volume sampler. No major immediate efforts should be made
to replace the current FRM; however, studies to improve and clarify the
existing FRM are strongly suggested. Exploratory studies to evaluate
the long-range possibility of replacing the hi-vol are also encouraged.
Local meteorology has a modifying effect on TSP levels. Moni-
toring network configuration and sampler placement can also significantly
-------
5
affect the measured TSP concentrations and our understanding of the
severity of the air pollution problem. Comparisons between cities and
rural a.r.eas using entirely different siting criteria may in many cases
be totally misleading. This can be a serious problem in high density
source areas emitting large dense particles. Such emissions result in
frequently changing steep concentration gradients both in the horizontal
and vertical direction. In such instances significant differences in
siting criteria will hinder the development of accurate and equitable
control strategies.
In addition to monitoring to determine national or regional
attainment or non-attainment of TSP ambient standards, a number of
other uses of monitoring data have been listed and reported. Table 1
lists some major uses of monitoring data but by no means is intended
to be all inclusive. Regardless of the specific monitoring purpose,
recognition and consideration must be given to the factors affecting
TSP levels discussed above.
SAMWG reviewed the monitoring systems operations for the variety
of aerometric data uses listed in Table 1. Monitoring for research
and enforcement case preparation purposes were not considered for dis-
cussion at this time. Four major issues were identified and are dis-
cussed later. The four issues are as follows: «
1. Should there be a change in the present monitoring networks
for total suspended particulates (TSP)?
2. Are improvements in the FRM for TSP needed?
-------
TABLE 1. Principal Uses of TSP Aerometric Data by Use Level•
EPA
USES HQ
1.
2.
3.
4.
5.
6.
7.
8.
9.
.0.
.1.
Judge attainment/non-attainment of TSP NMQS
Evaluate progress in achieving/maintaining
NAAQS or state standards
Develop or revise SIP's to attain/maintain
TSP NAAQS
New Source Review and prevention of signi-
ficant deterioration
Develop or revise national TSP control
policies (e.g., NSPS, tall stacks, SCS)
Model development and validation
Energy Suoply and Environmental
Coordination Act (ESECA)
Support enforcement actions
Public information (e.g., air quality indices)
Health research/establish standards
Develop or revise local control strategy
1
1
1
1
2
3
1
2
1
3
1
EPA STATE/LOCAL
RO AGENCIES
3
3
3
3
N/A
3
2
3
1
3
2
3
3
3
3
N/A
3
3
3
3
3
3
12. Determine specific cause of pollution in
an area 12 3
13. Determine nature of air pollution problem
in an area 33 3
*Use level refers to the detail, extent and frequency of reporting
of data needed by the user for the stated purpose.
1. Refers to a low level of detail, extent, and frequency of reporting
2. Refers to a moderate level of detail, extent, & frequency of reporting
3. Refers to a high level of detail, extent, and frequency of reporting
N/A - Not applicable
-------
7
3. Should the sampling frequency for TSP be fixed or varied?
4. What are the emission data and modeling needs for improving
TSP control planning technology?
-------
8
ISSUE 1: Should There Be a Change in the Present Monitoring Networks
For Suspended Particulate Matter?
AIR QUALITY SURVEILLANCE NETWORKS
The EPA regulation, CFR 40 Part 51 17, describes the essential
/•
elements of a SIP air quality surveillance system. It provides a
delineation of the measurement method, minimum frequency of sampling,
and minimum number of air monitoring sites for air quality control
regions. The specified minimum number of monitoring sites varies with
the population of the region. The regulation further states that at least
one monitoring site must be located in the area of estimated maximum
pollutant concentration. Additional information for designing an air
quality monitoring network is also provided in appropriate EPA guidance
documents.
State and local control agencies have used the 40 CFR 50.17
minimum requirement criteria for numbers of stations, the air guidance
documents for network design and instrument siting, plus the experience
and judgment of their professional staffs in designing and locating sus-
pended particulate monitoring reworks.
Based on 40 CFR 51.17, the minimum number of TSP monitoring
stations required nationally is approximately 1360. Data from the 1974
f
Monitoring and Air Quality Trends Report show a national total of 2004
TSP stations reporting a valid year of data and an additional 1784
stations reporting a minimum of three 24-hour high-volume samples. The
primary annual standard (75 yg/m3) was exceeded at 467 (23%) of the 2004
valid year stations, while the primary 24-hour standard (260
-------
9
was exceeded at 326 (8.6%) of the 3788 stations reporting minimal data.
EPA recently completed a trend analysis of stations having his-
torical continuity. TSP sites having four consecutive quarters in the
1971-72 period and in the 1973-75 period were selected for analysis.
A total of approximately 1800 stations met the trend station criteria.
In 1975 approximately 1380 (77%) of these trend stations were below
the primary annual standard (75 yg/m3); 1000 (56%) were below the
secondary annual standard guide (60 yg/m3); 660 (37%) stations were
below concentrations of 50 yg/m3. These figures based on comparisons
with the national air quality standards demonstrate that a significant
number of stations meeting the above criteria for trend analysis may
very well be put to better use or discontinued. It is SAMWG's opinion
that a thorough evaluation of current TSP networks should be done by
air pollution control agencies and proper consideration be given to
State and local air quality standards which are stricter than NAAQS.
Each station's monitoring purpose should be reviewed and if there is
no further need for the data generated, the station should be discon-
tinued or relocated to areas where monitoring data are required. SAMWG
reviewed the various needs for ambient air quality data at the National,
Regional, State, and local control agency levels. In order to satisfy
these data requirements, SAMWG suggests an air quality surveillance
network comprised of three types of stations: National Air Quality
Trend Stations (NAQTS), State Air Quality Trend Stations (SAQTS), and
Special Purpose Monitoring Stations (SPMS). SAMWG suggests that the
SIP air quality surveillance network principally consist of National
-------
10
and State air quality trend stations. The special purpose monitoring
stations would supplement the SIP monitoring network efforts in certain
SIP problem situations requiring additional monitoring. In addition,
the special purpose stations would be used to respond to the numerous
non-SIP problems frequently encountered by control agencies.
A. National Air Quality Trend Stations
To meet EPA's national monitoring objectives, it is SAMWG's
opinion that a core of permanent national air quality trend stations
operated by State and local control agencies be established. These
national air quality trend stations (NAQTS) would provide data for the
following uses:
1. Balanced national and regional overview of air quality
levels and patterns
2. Broad based assessments of national trends and progress
in achieving NAAQS
3. A high quality data base for developing and evaluating
national control policies
The number of stations and location of stations would be
arrived at through a cooperative effort between the Regions and State
and local agencies. A set of minimum criteria would be used as a
guide in making decisions on the number, general location of stations,
and shelter placement in order to provide a certain level of uniformity
throughout the Nation.
-------
11
At each NAQTS sampling frequency and calibration and operating
procedures would conform to all Federal Reference Method requirements.
Implementation of the Federal guidelines on quality assurance for TSP
is an essential operation of the permanent trend stations. SAMWG rec-
ommendations for a minimum quality assurance program are described in
Appendix A. Data collected at all NAQTS would be reported to EPA.
SAMWG believes the present National Air Surveillance Network
(NASN) stations should be subjected to the same critical review and
evaluation suggested for other existing TSP stations. If appropriate,
they should be discontinued or relocated to meet the NAQTS criteria
summarized on page 15. Since we do have an extensive base of historically
continuous NASN data, effort should be made to retain as many existing
NASN sites as possible. Added emphasis is provided for retaining good
NASN sites due to the special filter analysis currently being conducted
and the projected use for retrospective analysis in special problem areas.
B. State and Local Ambient Monitoring Stations
In addition to the operation of the NAQTS, SAMWG believes State
and local agency SIP ambient air monitoring data needs require the oper-
ation of permanent State and local ambient monitoring stations (SLAMS).
The major SIP ambient air data uses for these stations include the following:
1. Judge attainment/non-attainment of TSP NAAQS
2. Evaluate progress in achieving/maintaining NAAQS or state
standards
-------
12
3. Develop or revise SIP's to attain/maintain TSP NAAQS
4. New Source Review and prevention of significant deterioration
5. Support enforcement actions
6. Develop or revise local control strategy
It is envisioned that these sites would also provide data use-
ful for population exposure studies, more precise definition of air
quality concentration gradients, and initial documentation of special
TSP problem areas. These stations must use the FRM. The EPA guidelines
on network design and siting should be used in establishing the stations.
Likewise, adherence to established quality assurance practices should be
followed in the operation of the SLAMS. There will be no national EPA
reporting requirements, however, Regional Offices may have reporting
requirements ranging from raw data to summary reports.
The number of stations required to satisfy the SLAMS data needs
would, of course, vary extensively from one urbanized area to another.
SAMWG suggests that the final number of state air quality trend stations
be decided through cooperative Regional Office and State discussions.
C. Special Purpose Monitoring Stations
As seen from Table 1, there are numerous uses of air quality
data by State and local agencies not fully accounted for by the national
air quality trend stations or the state air quality trend stations.
Examples of these additional uses are listed below:
1. Model development and validation
2. Energy Supply and Environmental Coordination Act (ESECA)
3. Public information (e.g., air quality indices)
-------
13
4. Health research/establish standards
5. Determine specific cause of pollution in an area
It is recognized that NAQTS and SLAMS data could very nearly satisfy
some of the data uses listed. Likewise special purpose monitoring (SPM)
may well be required to fully meet SLAMS monitoring objectives.
Because of the very special nature of this type of station, the
location of stations, number of stations, sampling frequency, length of
study, and measurement method selection should be the prime responsibility
of the agency doing the monitoring. The FRM, if at all possible, should
be used in certain situations the TSP nonattainment problem may be
sufficiently complex to require particulate sizing sampling techniques
plus detailed chemical or physical analysis. In such instances, the
choice of sampling and analytical technique should definitely be made
by the monitoring agency.
Similar to the discussions of the NAQTS and SLAMS, the SAMWG
believes whenever possible EPA guidelines on network design and siting
should be used in setting up these special purpose monitoring stations.
Quality assurance practices should, of course, be followed whenever
possible. Data collected at these stations would not be routinely
reported to EPA; however, brief summary reports should be made available.
RECENTLY IDENTIFIED TSP SITING PROBLEMS
A recent study concerned with the assessment of the national par-
ticulate problem identified two TSP station siting problems occurring
in several cities which SAMWG believes requires additional attention.
-------
14
Problem one was the general absence of background, nonurban stations to
monitor incoming transported air pollutants. Problem two was the lack
of representative industrial monitoring sites in some industrial areas.
In areas with major industrial centers which substantially exceed
the standards, the absence of adequate data for remote nonurban areas
is not critical at this time. However, when the major control strategies
are implemented, and the improving air quality approaches the standards,
more precise air quality control planning will require TSP pollutant data
for incoming transported air masses. In certain other urban areas where
traditional sources are not a major problem but the standards are close
to being exceeded, there currently are planning problems because the
agencies do not have data relating the changes in air quality levels
resulting from the possible transport of incoming particulates through
remote, suburban and into urban areas. In other areas where secondary
pollutant concentrations are significant, their sources are unknown
and transport of pollutants may be significant, this deficiency of non-
urban data is even more of a problem.
The second problem—the lack of representative industrial area
monitoring or source-oriented monitoring sites in certain industrial
localities is principally a problem in areas along narrow riverside
plains. The industrial complexes located along the plains are varied
in nature, widely dispersed and, consequently, difficult to monitor.
Other industrial areas not monitored adequately are those located just
beyond city boundaries where monitoring by the local agency does not
occur.
-------
15
RECOMMENDATIONS .
1. National Air Quality Trend Stations
SAMWG recommends through a cooperative effort between Federal,
State, and local agencies, a minimum number of permanent national air
quality trend sites (NAQTS) be designated and operated indefinitely.
The data from these stations would be used to analyze and report
National, Regional, and local trends. These stations need not be
equivalent to the minimum number of stations currently calculated from
40 CFR 51.17, but a subgroup of the total.
Stations selected as-a NAQTS must use the Federal Reference
Method for suspended particulates and must implement EPA's quality
assurance program for suspended particulates. They should represent
sites which have at least three continuous years of valid data. The
stations must also be located according to the most recent EPA network
design and siting guidance documents. In addition, the selected stations
should meet any one or combination of the following criteria:
a. represent to the extent possible the maximum concen-
tration point of an urbanized area greater than 50,000 population.
The site should not be significantly affected by an individual source.
In areas not previously monitored, emission data and modeling should be
used to establish site location.
b. represent an urbanized area which has a population
greater than 50,000, is below the secondary standard, and has a high
potential for significant air quality degradation.
-------
16
c. represent to the extent possible areas of maximum
population density.
2. State and Local Ambient Monitoring Stations
a. SAMWG recommends, through a cooperative effort between
the Regional Offices and State agencies, the designation of State and
local ambient monitoring stations (SLAMS).
The number of stations in this category would be
decided by the States and Regional Offices. The SLAMS must use the FRM
and adhere to EPA guidelines on network design and siting, as well as
the quality assurance guidelines.
b. SAMWG recommends, in view of the recently identified
TSP station location deficiencies, a more comprehensive monitoring net-
work for TSP problem identification in widely dispersed industrial areas
and industrial areas frequently located just beyond local jurisdictional
boundaries. This may require increased cooperation between State and
local agencies.
c. SAMWG recommends the establishment of nonurban back-
ground stations to allow for the measurement of incoming transported
TSP pollutants in special proolem areas. This recommendation is also
in response to current TSP station location deficiencies.
3. Special Purpose Monitoring Stations
SAMWG recognizes the numerous State and local demands for
special purpose monitoring (SPM) data. Because of unique requirements,
SAMWG recommends State and local agencies have prime responsibility
-------
17
for deciding the number of such stations, location, sampling frequency,
length of study, and measurement method.
Whenever possible, the FRM should be used. In addition, EPA
guidelines on quality assurance and network design and siting should
be followed to the greatest extent.
4. SAMWG recommends the current SIP air quality surveillance
requirements (40 CFR 51.17) be modified to include:
a. criteria for establishment and operation of national
air quality trend stations, including quality assurance, number and
location of stations, siting specifications, sampling schedule, and
reporting requirements.
b. designation of state air quality trend stations
c. provisions which allow the State and local agencies
more flexibility to respond to the variety of special purpose monitoring
associated with the overall goal of attaining, and maintaining, ambient
air quality standards.
-------
18
ISSUE 2: Are Improvements in the FRM for TSP Needed?
The national ambient air quality standards for participate matter
require the Federal Reference Method (FRM) for measuring suspended
particulates. The method specifies use of the high-volume sampler, and
is described in the regulation published in the Federal Register. Speci-
fications are given for flow rates, flow measuring devices, calibration
equipment, and other items. The sampling shelter specifications are
also provided. Particle sizes as large as 100 ym can be collected by
the FRM; however, the actual size of the collected larger particles is
heavily influenced by wind speed. The method does not differentiate
between particle sizes, except that the particles must be small enough
to remain suspended long enough to reach the sampler. Numerous problems
have been identified with the FRM, some of which are described below.
It has been found that a high-volume sampler furnished with a clean
filter and passively exposed to the atmosphere for five days without
operating the motor may collect approximately 10 to 12 yg/m3 of particu-
late matter. Longer time periods of passive exposure have shown further
increases. Also, suggestions have been made that air exhausted from
the high-volume filter. Reports have been made of the directional bias
of the method due to the configuration of the sampler shelter. At
least one agency has modified the design of the shelter to include a
"stack-type" shape so as to avoid the problem of directional bias.
The FRM does not contain specifications for filters. It is widely
-------
19
recognized that filter problems, such as variations in tensile strength,
particle collection and retention efficiency, excessive holes in the
filter, and excessive concentrations of allowable elements and ions per
filter have been frequently encountered. These problems have resulted
in difficulties with TSP weight determinations and trace element and
ion analysis. The EPA has prepared specifications for high-volume sample
glass fiber filters and has had partial success in procuring acceptable
filters. SAMWG believes that NAQTS and SLAMS should only use filters
conforming to the EPA specifications.
SAMWG is also of the opinion that the curent FRM is too loosely
written in terms of shelter design, filter use, static operation, and
so forth. The consensus of the Work Group suggests the EPA critically
review the current FRM with the overall goal of tightening up all speci-
fications and improving the calibration and operation procedures.
The tape sampler paper referred to the need for alternative pro-
cedures for the tape sampler in emergency episode situations. The
high-volume sampler would appear to be a prime candidate procedure. EPA
supported an evaluation study of the high-volume method for determining
suspended particulates ove.1 short sampling times. Results of the study
indicated that a 4-hour sampling collection period followed by a 2-hour
period for sample equilibration in a controlled environment can produce
a precision of about 5.2 percent. Results of the study can be applied
with confidence only to the local area where the measurements were made.
Additional field evaluation under a variety of conditions is required
-------
20
to determine the applicable relationship for other areas. It is believed
that additional investigations of the use of the high-volume sampler for
emergency episodes is warranted.
As was mentioned in the introduction particle size monitoring is
of major concern for numerous reasons. In addition short-term or con-
tinuous particulate monitoring with provisions for chemical constituent
analysis is also of deep interest to many investigations. These measure-
ments are necessary to provide additional data for thorough evaluation
of the suspended particulate problem. In particular additional data of
this type are necessary to assist in assessing health effects, to deter-
mine short and long-range transport, to better understand non-attainment
problems and source receptor relationships and to more accurately deter-
mine particulate background levels. Such monitoring activities could be
carried out under the auspices of special project monitoring.
Recognition is given here to the problems encountered in siting
suspended particulate samplers. The issue, however, will not be dis-
cussed here but will be addressed in a separate siting issue paper.
RECOMMENDATIONS
1. EPA should critically review and evaluate the current FRM
writeup and tighten up the shelter design specification, prescribe filter
specifications, and attempt to clarify and improve the calibration and
operation procedures.
-------
21
r
2. EPA should explore the long-range possibility of replacing
the high-volume sampler with an effective and practical monitor capable
of providing particle sizing information and short-term or continuous
particulate monitoring capabilities.
3. All NAQTS and SLAMS must use filters conforming to the EPA
specifications. To the extent possible, the special purpose monitoring
stations should also operate with the standardized filters.
4. EPA should conduct further investigations of the feasibility
of the use of the high-volume sampler in emergency episode situations.
In addition other techniques for continuous monitoring of particulates
should be studied and considered for episode use.
-------
22
ISSUE 3: Should the Sampling Frequency for TSP be Varied or Fixed?
The optional sampling schedule for a specific TSP.monitoring
station depends upon a variety of factors, such as local emissions and
meteorology, the concentration range at the site* and, of course, the
intended purpose for the data. Possible sampling programs would include
the standard once every sixth day schedule, intensified sampling during
certain times of the year chosen on the basis of seasonality, or episode
monitoring to increase the probability of detecting peak values. Each
type of schedule has certain advantages depending upon the primary use
of the data. Therefore, to facilitate the discussion, the sampling
frequency is treated separately for the three types of sites: National
Air Quality Trend Stations (NAQTS), State and Local Ambient Monitoring
Stations (SLAMS), and Special Purpose Monitoring Stations (SPM). This
categorization provides a convenient framework for discussing sampling
frequency with respect to the primary intended use of the data. To
provide some perspective for this discussion, a brief treatment of
statistical considerations is included.
STATISTICAL CONSIDERATIONS
Assuming that the sampling schedule is intelligently designed, it
is a truism to say that more data means less uncertainty. The main
point is whether the increase in information is worth the additional
resources required to obtain more data. With this in mind, it is use-
ful to briefly consider certain statistical aspects of sampling frequency
Any air quality value involves variability associated with many
factors, such as meteorology, emissions, measurement methodology, or
pure sampling error. This discussion focuses only upon sampling error;
-------
23
that is, the error that could be eliminated by increased sampling.
For TSP there are both long- and short-term standards so not only is
the geometric mean of interest but also the peak values. The need for
both types of information complicates the choice of a sampling frequency.
Tables 2 and 3 serve to illustrate this point by indicating the approxi-
mate error in estimation the geometric mean and the chance of correctly
determining compliance with the once per year peak value standards. For
example, Table 2 indicates that with once every sixth day sampling, there
is a 90 percent certainty that the error in estimating the geometric
mean is less than 10 percent. In Table 3, it is shown that if a site
had 8 days above the standard, an every sixth day schedule would only
have a 40 percent chance of recording two or more violations. In other
words, there is a 60 percent chance that such a site would appear to be
in compliance. The 10 percent error in estimating the mean is probably
on the same order as the variation due to year-to-year fluctuations in
meteorology and is reasonably accurate for most sites.
The main advantage in increased sampling frequency is seen in Table
3 in terms of detecting peak values. However, if this is the primary
reason for increased sampling frequency, then an across-the-board
increase throughout the year is not necessarily the most efficient means
to obtain this information. There is no need to intensify sampling
during seasons when values are low if the primary purpose is to sample
for peak values. A more efficient approach is to increase sampling only
during peak seasons or to use episode monitoring. The specific option
chosen may well vary from site to site depending upon local factors.
-------
24
TABLE 2 . Approximate Sampling Error in Estimating the Geometric
Mean for TSP (SGD = 1.6)
Sampling Frequency
Certainty Every 6th day Every 3rd day Every other day
90% +10% + 6% + 4%
95% + 12% + 7% + 5%
TABLE 3. Probability of Selecting Two or More Days When Site
Exceeds Standard
number^of Sampling Frequency, days/year
excursions 61/365 122/365 183/365
2 0.03 0.11 0.25
4 0.13 0.41 0.69
6 0.26 0.65 0.89
8 0.40 0.81 0.96
10 0.52 0.90 0.99
12 0.62 0.95 0.99
14 0.71 0.97 0.99
16 0.78 0.98 0.99
18 0.83 0.99 0.99
20 0.87 0.99 0.99
22 0.91 0.99 0.99
24 0.93 0.99 0.99
26 0.95 0.99 0.99
-------
25
NATIONAL AIR QUALITY TREND STATIONS (NAQTS)
The National Air Quality Trend Stations are designated for several
purposes, but the primary goal is to provide timely high quality data
that can be used to assess trends and evaluate progress in achieving/
maintaining standards throughout the nation. A direct consequence of this
»
aim is the need for uniformity and rapid response. For the most part,
the data will be aggregated for clusters of sites and results will be
given for broad groups of similar types of sites. The emphasis on uni-
formity suggests that all of these sites use the same sampling schedule.
The once every sixth day schedule would be adequate for this purpose.
The uncertainty concerning high values could be compensated for by using
extrapolation techniques. Although these techniques have certain limi-
tations, they are sufficient for making inferences about general trends
for groups of sites. If a local agency needs additional sampling to
obtain more information concerning peak values or more precision for the
mean, then they can supplement the schedule by adding additional samples.
However, the most effective scheme for supplementing the basic schedule
may differ from one area to another. Therefore, it is essential to allow
flexibility in how the frequency may be increased. Any balanced schedule,
such as every sixth day or every third day, would be acceptable because
no weighting formulas or other special data manipulations would be
t
required in estimating means or percentiles. Any departures from a
balanced schedule would necessitate both special consideration in EPA's
data analysis and also more information from the local agency concerning
-------
26
the nature of these departures. For example, data obtained by episode
monitoring would have to be treated differently than data from a site
employing a predetermined fixed schedule with more intensive sampling in
peak seasons. Therefore, for data reporting to EPA from the NAQTS,
SAMWG believes a minimum balanced sampling schedule of once every sixth
day be required or any more frequent balanced schedule, such as once
every third day. Any additional data obtained from unbalanced schedules
that augment the basic once every sixth day schedule are not to be sub-
mitted to EPA. Thus, the local agency need not keep EPA informed of how
they are varying their basic balanced schedules but may still use these
additional samples in combination with the balanced schedule data sent
to EPA to evaluate compliance or progress at that site.
STATE AND LOCAL AMBIENT MONITORING STATIONS (SLAMS)
The main points in the previous discussion are also applicable for
SLAMS. The major differences are that there is much less separation
between the collector and the user of the data and that more precise
statements may be necessary concerning compliance at a specific site.
A flexible schedule may be t"j more easily and efficiently for these
purposes. For example, once every sixth day sampling may be used as
the basic schedule and augmented as needed. As shown in Table 1, if
the geometric mean is more than 10% below the standard, then ontfe in six
days is adequate. Figure 1 illustrates how the degree of uncertainty
decreases with additional sampling. If the primary concern is peak
values, then additional sampling may be needed, but this may be done
-------
27
85
3 80--
£75
65 L
1
' :
-
-
f
' *
:,>
- — A'Jove the standard
- — Area of uncertainty
- — Below the standard
n
,
!;- ~
'•
-
61 91 122 183
Number of sampling days per year
365
Figure 1. 55% confidence intervals about the annual primary standard for
TSP for various sampling frequencies (assume the standard geometric devia-
tion equals 1.6).
efficiently by increasing monitoring only during peak seasons. However,
as discussed earlier, the use of an unbalanced schedule requires more
care when computing means or percentiles or the results will be biased.
A routine weighting factor approach is adequate in some cases but is
further complicated if episode monitoring is used. Therefore, while
these modified schedules are quite good for assessing compliance with
peak value standards, they require additional care when the data are used
for trends to ensure that the data sets are comparable from one year to
the next.
SPECIAL PURPOSE MONITORING (SPM)
As the name indicates, these monitors are for special purposes and,
as would be expected, there are virtually no constraints on how the
sampling schedule is varied. The major concern is that the sampling
schedule be consistent with the intended use of the data so that any
inferences made are sufficiently accurate for the purpose in mind. The
list of possible options is almost endless, and those familiar with air
pollution data are well aware of the many possibilities. For example,
two data points may be sufficient to show a violation of the "once per
-------
28
year" standards. Data from only the peak season may be sufficient to
show compliance with the annual mean. All that is really required to
determine the adequacy of a sampling schedule for special purpose moni-
tors is that careful attention be given to how the data will be used
before even the first data point is collected.
RECOMMENDATIONS
On the basis of statistical considerations and the purposes of the
data, SAMWG recommends the following schedules for TSP monitoring:
1. National Air Quality Trend Stations employ an every sixth day
schedule as a minimum. Data obtained from any balanced schedule with
more frequent monitoring is also acceptable. Additional data that result
in an unbalanced schedule are not to be submitted to EPA but may be used
by State and local agencies in combination with the balanced schedule
data sent to EPA to evaluate compliance or progress at that site.
2. State and Local Ambient Monitoring Stations employ an every
sixth day schedule as a minimum. This basic schedule may be augmented
as required by the State or local agency to ensure sufficiently precise
statements regarding the status or progress of the station.
3. Special Purpose Monitoring stations may employ any monitoring
schedule that is consistent with the intended use of the data.
-------
29
ISSUE 4: What Are The Emission Data And Modeling feeds For Improving
TSP Control Planning Technology
The study "National Assessment of the Participate Problem" identified
several monitoring areas currently lacking information for sound air
quality planning decisions. Two of these topics were TSP air quality
simulation or dispersion models and additional detailed source emission
data for special problem areas. As described in other issue papers,
ambient air quality models are mathematical expressions of the relation-
ships between source emissions and measured ambient air pollutant con-
centrations. Sound control decisions based on atmospheric simulation or
dispersion models require comprehensive documentation of all major
sources and their emissions. The particulate assessment study indicated
a need in certain locations to identify and quantify TSP emissions from
non-traditional sources, natural and transported particulates, as well
as the generally well documented traditional sources. For TSP, fuaitive
emissions from traditional sources have generally been assumed to be
almost negligible in comparison to stack emissions. As an example, how-
ever, recent FPA findings revealed that fugitive particulate emissions
from the charging of electric arc furnaces can result in up to 50 times
the stack emissions released downstream of the control device. Also
fugitive emissions are usually released close to the ground resulting in
little dilution with the net effect of high concentrations at the ground
*
level.
3
Monitoring sites influenced by fugitive emissions averaged 25 vg/m
higher than industrial sites affected by stack emissions alone. It
-------
30
appears that traditional source fugitive emissions in certain indus-
trial areas of the country constitute a significant portion of the TSP
problem and will be an increasing problem in other areas. Effective
control of these sources seems to require proper understanding of the
intensity of emissions from these sourc<_o.
In many cities, the r^jor traditional sources of TSP, i.e., 50,100
tons/year, have significantly reduced their emissions to the point
where smaller sources of TSP are now the major concern if ambient stan-
dards are to be attained. In such areas additional emission inventories
for the smaller sources appear to be necessary for proper control deci-
sions.
It has also been found that cities having relatively insignificant
TSP emissions from traditional sources were still measuring TSP levels
30 vig/m3 higher than nonurban concentrations resulting in levels higher
than the secondary annual standard guide. Modeling efforts in such
areas using traditional source emission data failed to explain the
measured ambient concentrations. Particulate levels in these cities
were found to be caused by concentrated urban activities and these
activities have been designated as nontraditional sources. Two categories
are included under nontraditional sources. One category covers specific
sources such as construction and demolition activities, tailpipe emis-
sions and tire wear. The second category covers urban activities that
result in particulates becoming entrained or reentrained. Particulates,
man-made or natural, that become airborne because of wind or man's
activity are referred to as fugitive dust emissions. Included under
-------
31
fugitive dust emissions are windblown particulates from paved and
unpaved roads, from agricultural activities such as planning, discing
and cultivating, and from exposed surface areas at construction sites.
Fugitive dust emissions are not to be confused with fugitive emissions
from traditional industrial sources wntch are emitted to the atmosphere
from doors, windows, budding cracks, vents and so forth.
Efforts have been made to assess the national impact of these non-
traditional sources on ambient TSP levels at different site types. For
example, reentrained particulates at industrial monitoring sites can
add 15 to 35 yg/m3 to the TSP values. These levels are average values
and should be expected to vary from one area to another. Construction
and demolition activities impact locally on TSP levels but the effect
is not easily predicted. Increases in TSP loadings from these kinds of
activities are related to the type and level of activity, distance from
the activity, and so forth. Increases of 10 to 15 vig/m3 have been noted
at distances within half a mile of such activities. Current emission
factors, however, do not consider fallout and deposition of particles so
new emission factors should be developed. It appears that control
strategies for attainment of national standards in urban areas, where
violations of TSP standards are measured and believed to be caused by
nontraditional sources, will require detailed emission inventories of
these types of sources. If possible particle size as well as temporal
and spatial variability should be considered when developing emission
inventory techniques for these sources. It would also be helpful to
have a diffusion model which considers particle size, small scale dif-
fusion, fallout and deposition of particles.
-------
32
Assessment of the TSP problem also indicated that short and long-
range transport of natural, primary and secondary participates can signi-
ficantly affect ambient TSP levels. The discussion here will confine it-
self to the transport of particulates generated by man's activities.
Long-range transport of particular refers to the transfer of
particulates in an air mas^ over a distance of several hundred kilometers.
Short-range transport refers to transport over distances ranging from a
few up to 100 kilometers. Data from one study indicate the possible
transport of particulates over a total distance greater than 1200 kilo-
meters. The TSP concentrations over a six-day period increased from 68
to 204 yg/m3- In this case the moving air mass may have accumulated
particulates and sulfate precursors from industrial activities along the
path. Much further work is needed in this area to support any firm con-
clusions on possible transport effects. Estimates of transported pri-
mary particulates for eastern cities averaged about 5 yg/m3 but were
considered much higher in the densely populated northeast. The west and
midwest effects are thought to be minimal. Short-range transport of
particulates is a difficult problem to manage particularly when the trans-
ported particulates cross jurliuictional boundaries and the most effective
method of control cannot be agreed upon by adjoining agencies.
Transported secondary particulates are generally considered to
include sulfates, organics, and nitrates. They can be formed over a long
or short time period and can also undergo long- and short-range transport.
The sulfate levels in the long-range transport study mentioned above were
three times higher than the normal values. In the west and midwest trans-
ported secondary particulates are believed to contribute a minimal amount
-------
33
to local particulate concentrations. In the east, estimates of 10 yg/nv^
have been used.
Of additional concern are secondary participates formed in urban
areas from local sources of emissions. Proper control strategy planning
demands that secondary particulates fo.meti within the jurisdiction of
the control agency be considered separately from those transported into
the city. NASN data for sulfates and nitrates show the largest increase
in secondary particulates occur in heavily industrialized areas where
secondary particulate levels inside the urban area average about 6 ug/m3
above the nonurban values. Overall, cities may have secondary particu-
late levels 5 to 15 pg/m3 higher than nonurban values, due primarily to
sulfates and nitrates.
The above discussion suggests TSP levels in certain cities may have
been transported into the area from other jurisdictional areas. Once
again substantial additional studies are needed to support such suggestions.
The particulates may be primary or secondary pollutants generated a few
kilometers upwind or several hundred kilometers away. The short-range
transport problem may be adequately solved by cooperative efforts between
the adjoining agencies. Larger scale transport problems can be extremely
complicated as a result of chemical and physical transformations, removal
mechanisms, variable meteorological conditions and so on. Solutions to
these types of problems will require regional and a national scale
strategy. Continuing efforts are needed to acdurately determine the rate
and extent of formation of secondary particulate pollutants, principally
sulfates and nitrates. Data are also needed on factors influencing rates
-------
34
of formation of sulfates and nitrates as well as on the factors affecting
the chemical composition of sulfates and nitrates.
Emission monitoring data requirements and needs have been discussed
in the issue paper, "Strategies for Point Source Monitoring." The dis-
cussion included continuous in-stack monitoring and manual source
testing. Table 2 summarizes the continuous emission monitoring require-
ments by source category. The suggestions relating to particulate matter
source monitoring deserve repeating here. The Work Group recognizes the
need for data correlation between opacity monitoring and particulate
emissions. Many additional sources need to be included by the opacity
monitoring requirements. Finally, it is suggested that EPA establish
quality assurance procedures for calibration and operation of source
continuous monitors.
RECOMMENDATIONS
1. EPA should develop better techniques for estimating fugitive
emissions from traditional sources.
2. Improved inventories of fugitive emissions from traditional
sources should be conducted in heavily industrialized areas.
3. State and local agencies should reassess their point source
cut-off points in TSP problem areas to be certain that smaller sources
which may now constitute a major percentage of the traditional source
emissions, are receiving appropriate attention.
4. EPA should develop improved techniques for inventorying emissions
from nontraditional sources in urban areas and support their use by
State and local agencies.
-------
35
5. EPA should develop and provide to the States a diffusion model
which considers particle size, small scale diffusion, and the deposition
characteristics of emissions from nontraditional urban activity sources.
6. EPA should continue efforts to develop and document the
mechanism of formation and models for the prediction of the secondary
particulate pollutants sulfates and nitrates.
7. EPA should continue to conduct and support research efforts
directed to improve our knowledge and understanding of short and long-
range transported particulates.
8. EPA should continue its efforts to develop continuous monitoring
instruments to monitor particulate emissions from industrial sources.
The agency should also support studies to provide the appropriate data
necessary to convert opacity measurements made by existing continuous
monitors into particulate emission rates.
9. EPA should establish quality assurance procedures for calibration
and operation of source continuous monitors.
-------
36
Table 4
CONTINUOUS EMISSION MONITORING.
Source Category
fossil fuel fired
'steam generators
sulfuric acid
nitric acid
-catalyst regener-
ators for fluid bed
catalytic cracking
Units
electric arc
furnaces
'Pollutants
opacity
so2-
N0x"
°2 or C02
SO,
NO,
New
Existing
opacity
opaci ty
>250 million
Btu/hr
aTl sources
covered by
NSPS
all sources
covered by
NSPS
all sources
covered by
NSPS
all sources
covered by
NSPS
>250 million Btu/hr
S02 only If flue gas
desulfurization
NO only where control
• strategy required
09/C05
if state
only
has emission
regulations
>300 ton/day produc-
tion
>300 ton/day produc-
tion and only where
control strategy
required
>20,000 BBl/day .
no requirements
primary copper,
zinc and lead
shelters
opacity
SO-,
all sources
covered by
NSPS
no requireiTEnts
ferroalloy
production
facilities
opacity
all sources
covered by
NSPS
no requirements
-------
A-l
APPENDIX A
RECOMMENDATIONS FOR A MINIMUM QUALITY ASSURANCE PROGRAM
I. Recommendations for Air Pollution Control Agencies Conducting
Monitoring Activities.
The quality of reported measurement data is substantially controlled
by a few main quality assurance activities. These activities, including
a suggested minimum content and a sequence for development, are discussed
below. It is recommended that the following activities be developed
immediately in an agency quality assurance program.
A. Agency Quality Assurance Policy and Objective - Each agency should
develop a written quality assurance policy and this policy should be made
known to all agency personnel. At the minimum this policy should project
a consciousness for quality assurance activities, provide specific pro-
cedures for implementing a quality control program, provide for corrective
action when required, state quality assurance objectives for each major
monitoring project operated by the agency, and explicitly delegate authority
to implement quality assurance systems planned by management officials.
B. Organization and Responsibilities - An organization chart showing
the key agency personnel and their area of quality assurance responsibility
should be prepared. A Quality Assurance Coordinator for the agency should
be designated. This designee should be responsible for coordination of
quality assurance activities within the agency and with other agencies in-
cluding EPA.
-------
A-2
C. Items to Be Developed Within the First Six Months by An Agency
Implementing a Formal Quality Assurance Program.
1. Measurement Method Review and Application - All existing
methods (sampling and analysis) used for routine ambient air and source
emission measurements should be reviewed and revised if necessary; new
written procedures should be prepared where none exist. A document control
system should be developed for these methods to keep agency personnel
abreast of changes in methodology. Any ambient air monitoring for criteria
pollutants conducted under State Implementation Plans must use the EPA
reference methods or EPA approved equivalent methods. Those states that
have received the delegation of authority to enforce the Standards of
Performance of New Stationary Sources and the National Emission Standards
for* Hazardous Air Pollutants must use the EPA reference methods or EPA
approved alternate or equivalent methods when monitoring for these regulated
pollutants.
2. Calibration Procedure Review and Established Traceability
for Working Calibration Standards - Calibration procedures used for all
measurement methods should be reviewed, revised if necessary, documented,
and included in the method write up. Document control should also be
established for these calibration procedures to inform agency personnel of
any changes. As an agency policy, traceability of the accuracy of working
calibration standards should be established by comparing these standards
to standards of higher accuracy whenever standards of higher accuracy are
available.
-------
A-3
3. Intra-Agency Quality Control Procedures - The procedures used
during sampling and analysis to detect, correct, and record out-of-control
conditions should be defined.
4. Audit Procedures - Procedures should be selected and implemen-
ted that will permit comparison of the performance of the measurement system
(sampling and analysis) under routine operation versus an independent tech-
nique. Commonly, this independent technique is either a performance' audit
or the use of a dual measurement system. Results from these audit pro-
cedures are useful in detecting bias in the routine measurement system.
5. Inter!aboratory Testing - Each agency and its contractors
conducting monitoring activities should participate in the EPA quality
assurance performance surveys. These surveys are conducted by the Quality
Assurance Branch, Environmental Monitoring and Support Laboratory, located
at the Research Triangle Park, North Caroling (EMSL/RTP). Requests for
participation should be made at the EPA Regional Office.
D. Items to Be Developed Within the First 12 Months by An Agency
Implementing a Formal Quality Assurance Program.
1. Data Validation Procedures - Data validation procedures
should be prepared as soon as practical. These procedures should define
routine tests or checks performed on raw data during the validation pro-
cess.
2. Preventative Maintenance - A schedule for preventative
maintenance for each measurement method should be prepared identifying
the maintenance tasks and frequency required. Procedures for performing
the maintenance tasks should be documented.
-------
A-4
E. Other Items - There are many other areas which should be imple-
mented after the above mentioned activities are functioning routinely.
These areas, as well as the items discussed above, are described in
Volume I of the EPA, EMSL/RTP Quality Assurance Handbook which has been
widely distributed to regional, state, and local personnel.
II. Recommended Quality Assurance Activities for EPA.
The EPA Office of Research and Development should continue and expand
its role in developing the basic information and guidelines required in
implementing an air monitoring quality assurance program. EPA Regional
Offices should expand their roles in directing and assisting control
agencies in implementing quality assurance programs for ambient and source
emission measurement systems. These expanded efforts are necessary to
assure the quality of data used to make major environmental decisions
related to State Implementation Plans, Standards of Performance for New
Stationary Sources, and National Emission Standards for Hazardous Air
Pollutants and to set new air pollutant standards.
The EPA Regional Offices should establish and implement validation
procedures, based upon EPA quality assurance guidelines, to evaluate
the adequacy of ambient and source data collected and reported. In
cooperation with the Office of Air Quality Planning and Standards, ORD
•
will develop a program which will qualify air pollution measurement data,
from a quality assurance standpoint, as acceptable for the EPA data banks.
The EPA recommended minimum quality assurance program, although not
formally constituted as a regulation, should be regarded by EPA, state
-------
and local agencies as essential to the production of valid air pollution
measurements and should be made an integral part of the monitoring pro-
grams by each agency.
-------
SUMMARY OF COMMENTS ON STRATEGIES FOR IMPROVED
SUSPENDED PARTICULATE MONITORING
The suspended particulate monitoring issue paper was distributed
to State and local agencies, EPA Regional Offices, EPA Headquarters
and Program Offices and the Council on Environmental Quality (CEQ).
Forty-six reviewers submitted written comments to SAMWG. The
number of reviewers by agency or office is as follows: State agencies
(17), local agencies (16), EPA Regional Offices (9), EPA Headquarters
and Program Offices (3) and Council on Environmental Quality (1).
Nineteen of the reviewers were in general agreement with the issue
paper as it was presented. The remaining twenty-seven reviewers sub-
mitted one or more comments for consideration in the drafting of the
air monitoring strategy document and the writing of the final version
of the suspended particulate monitoring issue paper.
A summary of the reviewers major comments and SAMWG's reply
follows:
1. Fifteen respondents discussed or mentioned the need for moni-
toring methods which would be an improvement over the present FRM for
suspended particulate monitoring. Most of these commenters expressed
the need based on suspected health effects and visibility reductions,
for methods which would separate the suspended particulates into size
fractions. A few of the commenters disagreed with the issue paper
statement ". . .No practical procedures for determining equivalency
-------
2
of alternative methods to the participate FRM currently exist or are
anticipated in the near future. As a result, the high volume sampler
is the only currently acceptable method." Some reviewers felt the
current FRM is the least acceptable method. In contrast, several
reviewers specifically mentioned their agreement with SAMWG's concept
of not initiating major efforts to investigate procedures to demon-
strate equivalence with or replace the high volume sampler as the FRM
for suspended particulate matter. SAMWG's rationale for taking the
approach stated above was because the major thrust of SAMWG's review
of air monitoring operations was SIP oriented monitoring. Since the
current NAAQS for suspended particulate matter was based principally
on high volume sampler data, the FRM selected for suspended particulate
monitoring was the high volume sampler which of course is required for
SIP monitoring. In addition, the EPA program responsible for equiva-
lent and reference method designation foresee major difficulties in
developing equivalent testing procedures for suspended particulates
and do not anticipate any being developed in the near future. As a
result the only currently acceptable method for SIP suspended particu-
late monitoring is the high volume sampler.
SAMWG does agree, however, that there is a real need for
particle size monitoring and short-term or continuous particulate moni-
toring with provisions for chemical constituent analysis'. Several pro-
totypes and commercial instruments capable of meeting some of the stated
requirements currently exist. These measurements are necessary to
-------
3
provide additional data for thorough evaluation of the suspended par-
ticulate problem. In particular additional data of this type are
necessary to assist in assessing health effects, to determine short
and long-range transport, to better understand non-attainment problems
and source receptor relationships ana to more accurately determine
particulate background levels. Such monitoring activities could be
carried out under the auspices of special project monitoring.
Changes in the issue paper will be made to acknowledge the
continuing need for the special purpose particulate monitoring needs
summarized above. SAMWG will also suggest that the agency should
explore the long-range possibility of replacing the high volume sampler
with an effective and practical monitor capable of providing most of
the requirements mentioned above.
2. Four reviewers strongly disagreed with .the use of short-term
hi-vol sampling for emergency episode situations. Disagreement with
the use of the hi-vol was based on the following arguments: too many
samplers would be needed for continuous monitoring, significant number
of mandays would be required to deliver and retrieve samples, excessive
amount of mileage would be required to obtain area-wide coverage, and
substantial delays in obtaining concentration levels would be experienced;
Two of the above reviewers strongly defended the continued use
of the tape sampler for emergency episodes.
Several reviewers specifically expressed their agreement with
the use of the hi-vol in emergency episodes, considered it a step in
-------
4
the right direction and requested other techniques for continuous
monitoring of particulates be further studied and considered for
episode use.
By suggesting further study of the use of the hi-vol in
emergency episode situations SAMWG was 1,00 inferring that tape sampler
use in emergency situation* was not acceptable. The tape sampler
issue paper recommendations addressing the use of tape samplers in
emergency episodes still apply. SAMWG believes that short-term hi-vol
sampling has potential for use in certain areas even considering the
previous arguments opposing it and that a limited feasibility study is
warranted. SAMWG also concurs that other continuous particulate
analyzers be further studied and considered for episode use. This latter
suggestion will be incorporated into the issue paper.
3. Several agencies having urban areas affected by fugitive dust
emissions from agricultural activities pointed out the omission of a
discussion of this type of problem in the issue paper. The reviewers
suggested that additional studies be conducted in attempts to improve
the emission factors from agricultural activities as well as to improve
our knowledge of the short-ran^e transport of these type of emissions
into urban areas.
SAMWG agrees with these suggestions and will include them under
issue four of the issue paper.
4. One comrnenter pointed out a degree of inconsistency in SAMWG's
position regarding responsibility for special purpose monitoring (SPM).
He notes in one situation the statement is made: "Because of the very
special nature of this type of station, the location of stations, number
-------
of stations, sampling frequency, length of study, and measurement method
selection should be the prime responsibility of the agency doing the
monitoring." Under issue one recommendations, however, he points out
that SAMWG recommends the States have prime responsibility for deciding
number and location of SPM stations, sampling frequency, length of study,
and measurement method.
The reviewer strongly objects to the recommendation just stated.
It is their belief that the local agencies are in the best position to
determine resource allocations and the needs for conduct of special
purpose monitoring studies.
SAMWG agrees with this latter statement and will correct the
issue paper to reflect this approach and eliminate the paper's incon-
sistency.
5. One reviewer discussed SAMWG's failure to consider the fact
that certain States may have air quality standards which are stricter
than the NAAQS. Thus SAMWG's conclusions regarding the possible
elimination of some TSP stations because they are below the Federal
standards would be inappropriate for those stations which are below
the Federal standards but in fact are violating State air quality
standards.
Modifications will be made to the issue paper which acknowledges
the existence of State air quality standards and considers them when
reviewing air monitoring networks.
-------
6
6. Many commenters offered numerous suggestions for improving
the paper. Although the specific remarks will not be,discussed here,
the Work Group will include all corrections and appropriate suggestions
for improvement in the revised issue paper.
-------
COMPOSITION OF PARTICUIATE MATTER IN
CALIFORNIA HI-VOL SAMPLES*
by
SuzAnne Twiss
California Department of Health
Berkeley, California
*For presentation at the Third Interagency Symposium on Air Monitoring
Quality Assurance, Marriott Inn, Berkeley, California, May 18-19, 1977.
-------
INTRODUCTION
Since April 19?6, the California Air Resources Board Air Quality Network
stations have been collecting hi-vol samples on glass fiber filters. Samples
from approximately thirty of the stations are forwarded to the Air and
Industrial Hygiene Laboratory (AIHL) for lead, sulfate, nitrate and benzene
extractable organics analyses subsequent to the total mass determinations.
made by the Air Resources Board staff.
As most stations sample on a six-day schedule, approximately ^5 samples were
collected at each station between April and December 1976. This paper sum-
marizes data for nine of the stations throughout Northern and Southern
California.
ANALYTICAL METHODS
After weighing, the sample is mailed to AIHL folded within a glassine envelope
inside a numbered manila folder. The deposit area is measured and the filter
cut in half. One-half is extracted in benzene for the organics determination.1
The remaining half is split into two quadrants, one is extracted in water for
sulfate analysis by the turbidimetric method2 and nitrate determination by the
brucine method . Lead is determined by wavelength dispersive x-ray fluores-
cence spectrometry on a disk, 38 mm in diameter, punched from the remaining
quadrant.4 The quadrant with the punched-out disk is retained for historical
purposes, except for 10% which are reanalyzed for lead by atomic absorption5
to provide quality control for the x-ray system.
- 1 -
-------
The raw analytical data are key-punched for computer processing. Final re-
sults are calculated by a computer program which prepares quarterly reports
and data cards for the Air Resources Board.
RESULTS
Total Mass
Figure 1 shows frequency distributions for total mass at nine California
stations during the period April through December 19?6. The x-axis is con-
centration, and the y-axis the number of occurrences of samples in each
concentration interval. The stations are arranged roughly from north to
south from the top to the bottom of the page. Of interest is the frequency
of 2^-hour averages above the California standard, 100 /jg/m3 for 2.k hours,
indicated by the dotted line running through all the charts, and the occur-
rence of concentrations greater than 200 /jg/m3 at Fresno, Bakersfield and
Riverside.
Most of the distributions appear log normal and Riverside particularly so.
Unlike the other stations which sample on a six-day interval. Riverside
samples every three days so there are twice as many samples represented here.
This provides data for evaluating the adequacy of a six-day schedule. By
selecting every other sample from Riverside, one can simulate two alternating
six-day schedules and compare them. For example, measurements made on days 1,
U, 7? 10, 13 and 16 could be divided into two six-day schedules: Schedule A,
days 1, 7, 13 and Schedule B, days k, 10, 16.
This was done for the Riverside mass data and is shown in Figure 2. The A
and B schedule distributions appear quite different although the geometric
- 2 -
-------
Figure 1
TOTAL MASS
0
LAKEPORT
0
SACRAMENTO
REDWOOD CITY
BAKERSFIELD
SANTA BARBARA
*'$^
^ ' *
*!?
'* .
iv1
'•*?-
fef^
,,;'f>
^
S^
•^;%
SAN DIEGM
1 i i i i i
50
100 150
200
- 3 -
250 300 350 400
CONCENTRATION, jug/m3
-------
Figure 2
TOTAL MASS - RIVERSIDE
15,-
10
h-H
0
CO
a
I—H
SCHEDULE A
geom. mean 123.9jig/m3
SCHEDULE B
geomf mean 123.3jng/m
0
COMBINED
*
geom, mean 123.Bug/in*
50
100 150 200 250 300 350
CONCENTRATION, ug/m3
mean
7.
95% CONF.
INTV
-------
means are virtually identical (123.9 and 123-3 lUg/m3)- Thus, a six-day
schedule in Riverside to determine compliance with the California standard,
60 Mg/m3 annual geometric mean, would be adequate. However, when we look at
the number of days the California 2U-hour standard, 100 /jg/m3, is exceeded,
slightly different conclusions could be reached. In Schedule A, there are
3k days > 100 iug/m3 and in B, 29 days. However, this difference is not
statistically significant.*
Lead
The frequency distributions for lead in these samples are shown in Figure 3-
Again, high 2^-hour averages are observed in Fresno, Bakersfield and
Riverside, but Sacramento, Santa Barbara and San Diego also have a signifi-
cant number of days over 1.5 /ug/m3, the California standard for the monthly
mean. /
Riverside again shows the more classic log normal distribution and one can
look at the two alternate six-day schedules to determine if the schedules
result in the same conclusions with respect to compliance.
The monthly geometric means for each schedule are shown in Table 1. The
column entitled "combined" lists monthly means for all the samples. In
each month, the geometric mean is above the California standard and the
level increases in the winter months.
*Arcsin transformation and test for difference in proportions, i.e. 3i4-/U3 vs,
29A3-
- 5 -
-------
Figure 3
LEAD
40-,
0
w 20-f
w
I
20-,
0-L
pj
S 20n
0-
20-,
0
LAKEPORT
SACRAMENTO
:„' •&>.
\ "• ;
- ''-H
'*S?*0>\
B A T T
SAL I
FRESNO
BAKERSFIELD
SANTA BARBARA
RIVERSIDE
SAN DIEGO
567+
CONCENTRATION, Mg/m3
- 6 -
-------
TABLE 1
LEAD - RIVERSIDE
MONTHLY GEOMETRIC MEANS
ug/m3
Month Combined Schedule A Schedule B
April 1.6 1.2 2.1 **
May 1.8 1.8 1.8
June 2.1 2.0 2.1
July 2.0 2.1 2.0
August 2.0 2.0 . 2.0
September 2.1 2.0 2.3
October 2.8 3.4 2.3 *
November 3.7 3.9 3.5
December 4.1 3.6 4.6
Schedules A and B significantly different at
** 95%
* 90%
by the Wilcoxon 2-sample test.
- 7 -
-------
The same seasonal variation is apparent in Schedule A and Schedule B.
However, in April, Schedule A indicates compliance, and Schedule B does
not. These means are significantly different at the 95$ level.*
Though total mass was not different for the two schedules, the lead data
indicate that other chemical species may be. Thus, a six-day schedule,
which is sufficient for determining compliance for a certain chemical
species, may not be adequate for all.
Of further interest, is the occurrence of very high lead concentrations,
particularly in Fresno and Bakersfield, where on several days, the lead was
higher than 7 Ag/m3. These are the same days, in winter, when the mass was
over 250 Mg/m3. This suggests the high mass concentrations observed in the
lower San Joaquin Valley may not be caused by wind-blown dust alone.
Sulfate and Nitrate and Organics
Sulfate is shown in Figure k. Bakersfield and Riverside have several days
in excess of 25 /ug/m3, the California standard. The highest sulfate con-
centration, 38 /ug/m3, occurred November 20 when the mass was 150 /jg/m3 in
Bakersfield.
Similar patterns are evident for nitrate, Figure 5. Fresno, Bakersfield
and Riverside had high nitrate occurrences with several days over ^5 /jg/m3.
On two days in December, the nitrate concentration in Bakersfield was more
than 80 iug/m3 when the mass was over 350 /^g/m3.
*Wilcoxon two-sample test.
- 8 -
-------
40n
as 20-1
20 -,
40 -i
10 •-,
0
0
Figure 4
S U L F A T E
SANTA BARBARA
LAKEPORT
SACRAMENTO
SALINAS
FRESNO
BAKERSFIELD
10 15
20
- 9 -
25 30 35 40
CONCENTRATION, jug/m3
-------
Figure 5
NITRATE
LAKEPORT
20-,
0-
SACRAMENTO
GO
w
—H
a
20-1
20-,
0
20-1
o_EE!Z
20-,
0
0
SALINAS
FRESNO
^^^ **V*'«
., " *. ''"'3n y%^*i > &-\
--.'"< -1 '"]
SANTA BARBARA
RIVERSIDE
:?a
SAN DIEGO
""*^>V';T^'^*1
10 15 20 25 30 35 40 45 +
CONCENTRATION, jug/m3
-------
Figure 6 demonstrates that Fresno and Bakersfield also experience high con-
centrations of organics, but Sacramento and Redwood City too have days in
excess of 25
Combined Species
The simultaneous occurrence of high mass with high lead, sulfate, nitrate
and organics suggests that these four species may comprise a fairly constant
proportion of the particulate matter present.
In Figure 7, lead, sulfate, nitrate and organics are plotted as a percent
.£
of the total mass for eight of the nine stations. From left to right
across the chart, the stations are roughly arranged in north to south order,
and from lower to higher population density.
There is a general trend, from north to south, for these four species to
account for an increasing proportion of the mass. There are no big dif-
ferences between stations; Salinas has proportionately fewer organics and
Riverside somewhat more nitrates, but at all of these stations these four
pollutants account for only 20-30$ of the total mass.
Of further interest is the fact that at Bakersfield, where the mass averaged
152 jUg/m3 for this nine-month period, these four chemical species account
for a similar percent of the mass as in Santa Barbara and Salinas, where the
average mass was 65 and 5^ /ug/m3, respectively.
*Redwood City is excluded because only organic and total mass data were
available.
- 11 -
-------
Figure t>
ORGANICS
40-i
30H
0-
20-,
CO
3 40n
«
0-
10J
0-
10H
30 -,
0
0
LAKEPORT
SACRAMENTO
REDWOOD CITY
SALINAS
FRESNO
BAKERSFIELD
•'" 'Vv. -
't'f',>-, '" x
v"\ '"'<•'' i
;^f
SANTA BARBARj
RIVERSIDE
- 12 -
30 35 40+
CONCENTRATION, jug/m3
-------
50
40
CO
30
20
OH
10
0
Figure 7
LEAD, SULFATE, NITRATE AND ORGANICS
AS A PERCENT OF TOTAL MASS
CALIFORNIA, APRIL-DECEMBER 1976
Lakeport Sacra- Salinas Fresno Bakers- River- Santa
mento field side Barbara
LEAD
S04
NO,
ORG.
San
Diego
-------
Size-segregated sampling at these locations would provide valuable information.
For example, if the sulfate and nitrate were predominantly present in large
particles, one could hypothesize wind-blown fertilizer to be the source. If
lead, nitrate and sulfate occur mainly in small particles, one could not
dismiss the Bakersfield air pollution problem so lightly.
In fact, several mechanisms probably produce high mass days in Bakersfield,
Without the particle size information, much of our interpretation is conjecture
Some further interpretation is provided in Figure 8. The upper half of this
chart is a day-by-day plot of lead, sulfate, nitrate and organics as a
percent of total mass. The dotted line is the average for this period. Note
that values below the mean tend to occur in the spring and summer, and values
above the mean in the winter quarter, October through December. The bottom
half of the chart is the day-by-day plot of total mass. Three peaks stand
out, May 6, November 2 and December 2.
On May 6, the mass was high and the four pollutants accounted for less than
10$ of the mass. This suggests a wind-blown dust situation and the meteoro-
logical data support this. There were high winds, up to 18 mph, preceding an
evening rainstorm. On November 2 and December 2, a different pattern was
observed. The mass was high again, but the four pollutants accounted for more
than the average proportion of the mass. These were calm, hazy days, in
winter, when the inversion layer is typically low. This suggests ordinary
air pollution episodes in Bakersfield.
It is evident from the data presented that some high mass days may be more
hazardous to health than others. Size-segregated sampling would allow us to
quantify this conclusion..
- Ill- -
-------
Figure 8
BAKERSFIELD
CO
CO
H
VJ1
CO
•5
to
50
40
30
20
10
0
350
300
250
200
150
100
50
0
Pb+S04+N03+ORG,
MASS
April
May
June
July
August
Sept.
October INovember
December
-------
REFERENCES
1. AIHL Method 6?, Determination of Total Organic Materials in Atmospheric
Particulate Matter, Air and Industrial Hygiene Laboratory, California
Department of Health, Berkeley, CA, April 1975.
2. AIHL Method 6l, Determination of Sulfate in High-Volume Particulate
Samples: Turbidimetric Barium Sulfate Method, Air and Industrial
Hygiene Laboratory, California Department of Health, Berkeley, CA,
July 1976.
3. AIHL Method 66, Determination of Nitrates in Atmospheric Particulate
Matter (Brucine Method), Air and Industrial Hygiene Laboratory, California
Department of Health, Berkeley, CA, September 1975-
k. Moore, H. , Application of Wavelength Dispersive X-Ray Fluorescence
Spectrometry to the Determination of Lead in Atmospheric Particulate
Matter Collected on High- Volume Glass Fiber Filters, AIHL Report 183,
Air and Industrial Hygiene Laboratory, California Department of Health,
Berkeley, CA, June 1976.
5. AIHL Method 5U, Analysis for Lead Content of Atmospheric Particulate
Matter Collected on High- Volume Glass Fiber Filters, Air and Industrial
Hygiene Laboratory, California Department of Health, Berkeley, CA,
December 197^-
- 16 -
-------
ACKNOWLEDGMENTS
I appreciate the technical and editorial assistance of Jerome Wesolowski,
Bruce Appel, and Walter John of the Air and Industrial Hygiene Laboratory,
and Spencer Duckworth and Robert Kuhlman of the California Air Resources
Board.
Hezekiah Moore, Russell Umbraco, and Aurora del Rosario supervised the
chemical analysis of the samples at AIHL, and Robert Snook, Victor Lee,
and Marlyn Garfinkel of that laboratory extended considerable effort in
the reduction and summarization of the data.
- 17 -
-------
SOUTH COAST AIR QUALITY MANAGEMENT DISTRICT
9420 Telstar Avenue, El Monte, California 91731
A
TECHNICAL SERVICES DIVISION
REPORT
A SIZE-DISTRIBUTION STUDY OF PARTICULATE MATTER
IN LOS ANGELES AND ANAHEIM
Air Quality Report No. 85
Margil W. Wadley, Supervising A. P. Chemist III
Metropolitan Zone
Robert D. MacPhee, Head, Chemistry Laboratories
Metropolitan Zone
William G. Bope, Chief APC Technical Services
Southern Zone
March 1977
John L. MiHs
Director of Technical Services
For presentation at the Third Interagency Symposium on Air Monitoring Quality
Assurance, Marriott Inn, Berkeley, California, May 18 and 19, 1977.
-------
PREFACE
Although both the federal and State governments have established air
quality standards for total suspended particulate matter (TSP), little
is really known concerning the variations in particle size distribu-
tion in the typical photochemical smog of this area. While it is
known that TSP levels here frequently exceed the 2^-hour standards
and, to date, invariably exceed the annual geometric mean standards,
it must be recalled that the specified high-volume measurement method
fails to discriminate between respirable and non-respirable fractions.
To refine our knowledge to the degree that new standards could be
written, identifying the dangerous fractions of TSP and their con-
stituents and directing control efforts mainly toward them, would
require a very large and thorough aerosol sampling and study pro-
gram. A really adequate program of this type would require resources
far beyond those available locally.
In order to gain some insight into the magnitude of this District's
TSP problem concerning respirable particles, however, a series of
fractionated aerosol samples was taken at Anaheim and in downtown
Los Angeles. This report presents the findings of this study over
the period of July I 97^ through June 1976.
70D107
-i-
-------
TABLE OF CONTENTS
Page
Preface . . i
Summary 1
Introduction .. 2
Experimental Procedures k
Results and Discussion 6
-i i-
-------
LIST OF TABLES
Table Page
I Monthly, Semi-Annual and Biennial Geo-
metric Means of Samples Taken by Ander-
sen Hi-Vol Particle Fractionating
Cascade Impactor and Regular Hi-Vol
Samplers 7
II Suspended Particulate Matter on Each
Stage of the Andersen Cascade Impactor
Expressed as Percentage of Total Mass
Col lected—1975 Calendar Year 8
-111-
-------
LIST OF ILLUSTRATIONS
Figure Page
I Typical log-normal plot of monthly geometric mean
data from the Anaheim station illustrates how the
overall mass median diameters (MMED's) were obtained
for the monthly geometric mean Andersen impactor com-
pilation in Table I 14
2 Geometric means for quarterly periods show the
difference in mass median effective diameters
(MMED's) between the stations 15
3 Monthly geometric mean mass loadings for Andersen
impactor samples show a general tracking of parti-
culate highs and lows for the three stations with
no clear seasonal patterns |6
k The mass-size distribution of total suspended par-
ti cu I ates for downtown Los Angeles and Anaheim are
plotted using the mass-density function on the or-
dinate 17
-IV-
-------
SUMMARY
This report presents the results of a two-year particle size study on
ambient air at two locations in the South Coast Air Basin (SCAB) using
Andersen high-volume cascade impactors together with comparison TSP
results using the common hi-vol samplers. The frequency of sampling
was once every six days except for the first 6 months of the study,
when it was every fourth or fifth day depending on location. The
chemical composition of the various fractions is not a part of this
paper.
More than 300 fractionated samples were collected and weighed at two
sites: two levels (22- and 85-foot) at the downtown Los Angeles office
of the District and one level (15-foot) at the Anaheim office of the
District.
In general, the total mass loadings (or TSP) were quite similar for
both methods at both locations. The geometric mean TSP for the two
years at the 85-foot level in Los Angeles was 107 jzg/m3 by combined
Andersen stages and 102 iig/m3 by hi-vol. No comparison hi-vol sam-
plers were run at the 22-foot level in this study. At Anaheim the
values were 12^4 jig/m3 for the combined Andersen stages and 102 /ig/m3
for the hi-vol.
The mass median effective diameter (MMED) of the particles was no-
ticeably smaller at both levels in Los Angeles (1.6 and 2.2 fjm)
than at Anaheim (2.7 /im). Furthermore, the MMED was smaller at
the 85-foot level than at the 22-foot level in Los Angeles. This
study showed that approximately 50 to 60% of the total particulates
at Anaheim and Los Angeles are in the respirable range, i.e., less
than 3.3 /*m in diameter.
-------
INTRODUCTION
The irritating properties of Los Angeles smog were recognized in the early
19^0's, and in the I950's it was recognized that the airborne participate
pollutants in the atmosphere of the South Coast Air Basin presented a po-
tential health hazard. Limited periodic surveys of total mass loadings of
particulate matter were conducted by one of the predecessors of this Dis-
trict (the former Los Angeles County Air Pollution Control District) in
I 9^9-1951 and by the Air Pollution Foundation in 1954. Since 195^ the
National Air Surveillance Network (NASN) of the U. S. Public Health Ser-
vice (USPHS) and its successor, the Environmental Protection Agency (EPA),
collected approximately 26 samples annually at several locations in this
Basin.
Measurements of total suspended particulate matter (TSP) by this District
using calibrated hi-vol samplers have been conducted at least once per
week on a regular continuing basis at not less than four sites in the Metro-
politan Zone of the Basin (Los Angeles County) since I965« Since September
1970, these measurements have been concurrent at all sampling locations for
the Z'f-hour period, midnight to midnight, so that the resultant data could
be assigned to a single calendar day. During the period 1970-7^, the number
of sampling locations progressively increased to eight and the sampling fre-
quency was changed to once every five days at all locations (July 1971). In
I975> sampling frequency changed to once every six days in accordance with
the new State ARB-coordinated schedule. The particulate filters thus ob-
tained in the Metropolitan Zone have been routinely analyzed for sulfate
and nitrate contents as well as for concentrations of seven metals. The
results of those analyses are presented in a Ten-Year Summary1.
Collection and analyses of TSP in the Southern Zone (Orange County) were
initiated in 1969 at one site in Anaheim, and the program has been expanded
progressively to eight locations. All samples were collected every fourth
day through I 97^. Beginning in I975, all samples have been collected on
the State-coordinated six-day schedule2.
This report describes additional and supplemental information over and above
that gathered in the two previously-cited programs. It covers Andersen im-
pactor sampling results over a two-year period (July 197^ - July 1976) at
Los Angeles and Anaheim. This special sampling was concurrent with the
regular hi-vol samplers using the Federal Register procedure3.
The Andersen hi-vol cascade impactors were operated at 20 cubic feet per
minute and collected TSP in five separate aerodynamic size ranges—up to
I.I f/m (backup), I.I to 2.0 f,im, 2.0 to 3.3 ion, 3«3 to 7.0 jjm, and larger
than 7.0 j/m.
-2-
-------
The true accuracy of the size separations isn't known with certainty. Work
by Burton* and a recent private communication provided by Burton to Lee5
indicate that Andersen fractionators can be used to determine with reason-
able dependability (may be slightly inaccurate in the smaller particle-size
ranges) the aerodynamic size distribution of suspended particulate matter.
Aerodynamic size means that, based on impactor theory where separation is
a function of both size and weight, particles behave as if they were dis-
crete spheres of unit density. Mass median effective diameter (MMED) dis-
cussed in the text to follow, means diameter of a sphere effectively be-
having as a unit density particle of that size. In 197^, hi-vol Andersen
cascade impactors were used in a nine-month study in Chicago*,and the size
separations reported seem, on an overalI average, to corroborate the data
reported here for Los Angeles and Anaheim. This information gives some
support to the idea that not only are particulates strangely similar in
two large cities but that the Andersen samplers seem to provide consistent
data and are relatively simple to operate in a network of stations.
Since the main reason for measuring particle size is the possible health
implication, it is appropriate to mention what the term "respirable frac-
tion*1 means. It is generally considered that particles larger than 10 ^m
are essentially all removed in the nasal chamber and have little possibility
of penetrating to the lungs. The efficiency of particle removal is high
in the pulmonary air spaces, being essentially 100% at 2 /jm, and falling
off to a minimum at about 0.5 fim. It then increases again as the force
of precipitation by diffusion increases with further reduction in size.
Therefore, those particles less than 3-5 /im in diameter are generally con-
sidered the respirable fraction7.
-3-
-------
EXPERIMENTAL PROCEDURES
Equipment and Supplies
Andersen Samplers. Hi-volume, multi-jet, multi-stage particle size
classifier, Andersen Model GMWL 2000H in aluminum shelter. Andersen
2000, Inc., Atlanta, GA.
Hi-Vol Samplers. Staplex hi-vols, Accu-Vol samplers and GMW hi-vols
were operated at different periods (see text under Procedural Details)!
Staplex Samplair, The Staplex Co., Brooklyn, NY; equipped with
8" x 10" (20- x 25-cm) filter holders (wooden shelters, flat roof),
GMW Accu-Vol, General Metal Works, Inc., Cleves, OH.
GMW 2000 Hi-Vol, General Metals Works, Inc., Cleves, OH.
The latter two samplers were in aluminum shelters with gabled roof.
They were calibrated with a BGI, Inc. variable resistance calibrator,
Model VRC, which was in turn checked for accuracy with a Roots posi-
tive displacement meter, Model 5MI25.
Air Pollution Balance. Torbal Model EA-IAP, The Torsion Balance Co.,
Clifton, N. J.
Temperature - Humidity Chamber. Blue M, Model FR-25ICX, Blue M Elec-
tric Co., Blue Island, IL. Used for conditioning filters in Metro Zone.
Storage for 2k hours over silica gel used in Southern Zone.
Glass Fiber Filters. The four upper stages of the cascade impactor
utilized commercially-available, perforated circular filters, Gelman
Type A, at all locations and for the entire period of the study. The
back-up filters (20- x 25-cm) on the impactor (last stage) were
matched to the type used on the adjacent hi-volume samplers at the
two locations. At Anaheim, these were Gelman Type A from July I 97*t
through December I97^» and Gelman Spectro Grade thereafter. At Los
Angeles, these were Gelman Type A from July 197^ through March 1975;
Spectro Grade from April through June 1975; and Gelman Type A/E there-
after. All filters were purchased from the Gelman Instrument Co.,
Ann Arbor, MI.
Air Flow Calibrat ion. The impactors were calibrated using an oil
manometer to measure pressure differential across the head. The
hi-vol sampler flow rates were measured using rotameters (Visi-Float)
which had been calibrated against orifice plates (Model GMW-25 Cali-
bration Kit, General Metal Works, Cleves, OH).
-k-
-------
Procedural Details For Collectinq Samples
The location of the samplers at Los Angeles was on the sixth-floor roof
(approximately 85 feet above the ground) and on the first-floor roof (ap-
proximately 22 feet above the ground) at the Metropolitan Zone headquarters
building in downtown Los Angeles. The samplers at Anaheim (lOlO S. Harbor
Blvd.) were on the roof of a maintenance building approximately 15 feet
above the ground. The flow rate for the Andersen impactors was 20 cfm as
specified by the manufacturer. The regular hi-vols were run at kQ cfm at
Los Angeles and 50 cfm at Anaheim.
A pair of regular hi-vol samplers was run in parallel with the Andersen at
the 6th floor roof in Los Angeles and a single hi-vol was run at the. Anaheim
location. These samplers are a part of the District's regular TSP network.
There was no regular hi-vol run on the first-floor roof in Los Angeles in
conjunction with the Andersen impactor located there.
Staplex hi-vols (without flow controller) were utilized at Los Angeles from
July through December 197^ and GMW Accu-Vol samplers thereafter. One GMW
hi-vol was run at Anaheim throughout the study (the latter had no flow
controller but a flow recorder). From July through December 197^, all sam-
ples at each location were collected concurrently from midnight to midnight;
however, the dates and frequency of sampling were different at Los Angeles
and Anaheim as described in the Introduction. Thereafter, alI samples at
all locations were concurrently collected every six days (using the State-
coordinated schedule, from midnight to midnight) unless failures occurred.
Subsequent to sample collection, the glass fiber filters from all stages
were conditioned for constant humidity prior to weighing. Previous work
has shown that such conditioning prior to sampling is unnecessary if fil-
ters are given reasonable care and handling.
-5—
-------
RESULTS AND DISCUSSION
A summary of total suspended particulate (TSP) mass concentration values in
terms of monthly, semiannual and biennial geometric means is presented in
Table I. Included also in the table are selected fractionation results and
mass median effective diameters (MMED). This table is a compilation pre-
pared from individual samples and fractionations which were too voluminous
to present in total.
Andersen fractionation results for all stages are reported on a monthly
geometric mean basis in Table II. These results have been presented for
the calendar year 1975 because of the large volume of data. The calendar
year presentation is sufficient to illustrate the method and draw the neces-
sary conclusions.
During the first twelve months of sampling at Anaheim, kk impactor (2^-hour)
samples were collected. Sixty such samples were collected at the sixth
fioor roof location at Los Angeles, and 5^ samples were collected on the
first floor roof. During the last twelve months, 56 samples were collected
at each of the two Los Angeles locations and 58 samples were collected at
Anaheim. The numbersof 2^-hour samples thus collected were approximately
equal at all locations.
Comparison of TSP by Andersen Impactor and Hi-Vol
An examination of the TSP geometric averages in Table I for the entire two-
year period shows that the Andersen impactor loadings exceeded the high-vol
loadings by about 5% at the Los Angeles 85-foot site. This difference
could be due to the larger number of filters in the Andersen device as op-
posed to the single one on the hi-vols, with the result that additional
artifact particulate mass was formed from gaseous air pollutants,but the
overall average TSP by Andersen impactor exceeded that by hi-vol at Anaheim
by 22%. These differences may also be partly due to hi-vol flow rates which
were not identical at the two locations. Prior work has shown an inverse
dependence of mass loadings on flow rate8. A recent Chicago study using
Andersen samplers showed the TSP results to exceed those from hi-vols by
$.k% at flow rates of 50 to 55 cfrn*.
Least-squares regressions (y = mx + b) were performed for the impactor/hi-
vol TSP values run side-by-side at Anaheim and Los Angeles. These compari-
sons were made over the twelve-month segments of data as well as for the
overall twenty-four month compilation (note: there were no comparative hi-
vol samples collected on the first floor roof at Los Angeles). The hi-vol
(HV) values were considered the independent variables and the impactor (IMP)
values were the dependent variables for the regressions summarized on Page 9.
-6-
-------
Table I
MONTHLY, SEMI-ANNUAL AND BIENNIAL GEOMETRIC
MEANS OF SAMPLES TAKEN BY ANDERSEN HI-VOL PARTICLE
FRACTIONATING CASCADE IMPACTORS AND REGULAR HI-VOL SAMPLERS
Month
and
Year
Jul . 74
Aug. Ik
Sep. 74
Oct. Ik
Nov. 74
Dec. Ik
6 Mo. GM
Jan. 75
Feb. 75
Mar. 75
Apr. 75
Los Angeles - Metropolitan Zone Laboratory
Sixth Floor Roof
(85 feet above ground)
Single Cascade Impactor
(20 CFM)
No. of
Sampl es
2
6
6
6
6
6
5
l»
6
k
May 75 I k
June 75 5
6 Mo. GM
July 75
Aug. 75
Sep. 75
5
5
3
% of Mass
< 1 .1 |im
38.4
31.3
36.6
k] .9
41.3
39.7
38.0
55.9
42.5
57.1
36.3
42.9
33.9
43.9
% of Mass
< 3.3 nm
46.4
54.9
56.6
57.3
58.2
51.4
54.1
69.0
58.0
70.1
55.2
62.3
56.0
60.2
40.9 61.0
38.5 60.5
42.9
Oct. 75 ! k j 33.8
Nov. 75 i 5 40.7
Dec. 75 : k 44.6
6 Mo. GM
Jan. 76 5
Feb. 76
Mar. 76
: it
40.1
66.5
50.0
58.8
63.1
59.7
46.0 62.1
43.0 60.2
6 35.5
53.5
Apr. 76 5 43.9 61 .2
May 76 5 1 30.1 51.0
June 76 5 | 32.1 .' 54 .k
6 Mo. GM : 37.7 56.2
24 Mo. GM | 39.9 57.9
MMEDa)
(urn)
1.6
2.4
2.1
I -3
1.6
i -5
1.7
0.6
1 .7
0.3
2.6
I .6
2.2
1 .2
ISP
fig/m3
85
^S
\ 1 7
1 ' /
162
i 18
i 16
92
M7
121
87
86
67
143
104
98
1.8 ! in
1.6 i 108
1.2 108
3.1 93
1.8 I'tO
1.4 152
1.7
1.4
1.6
2.1
1.4
2.9
2.7
I 17
149
83
91
86
101
119
1.9 !' 103
1.6 ; 107
Average of
Two Hi-Vol
Samplers
(40 CFM)
TSP ng/m3
82
93
148
99
121
94
107
104
76
88
64
118
110
92
111
90
99
93
145
162
114
168
86
87
82
99
108
102
102
First Floor Roof
(22 feet above ground)
Single Cascade Impactor
(20 CFM)
% of Mass
< 1.1 ^m
40.5
b)
30.5
31.6
38.1
35.5
34.2
40.5
28.5
24.9
42.0
38.6
32.0
33.2
39.8
32.4
37.1
32.0
39.6
44.1
37.2
48.7
47.6
30.3
30.0
26.8
31.8
34.3
34.7
% of Mass
< 3.3 nm
61.0
54.2
52.6
57.1
53.0
54.5
54.7
49.5
44.8
58.1
54.6
52.3
50.8
58.8
54.4
56.0
49.3
55.6
60.6
55.3
61.5
60.3
50.4
52.1
48.0
51.5
52.1
54.2
(ffn)
1.4
2.5
3.4
2.2
2.6
2.3
1.8
3.3
3.4
1.5
1.3
2.7
2.2
1.8
2.5
2.1
3.3
2.1
1.6
2.2
0.9
0.8
2.9
2.9
3-4
3.1
2.0
2.2
TSP
lig/m3
101
157
104
124
102
118
120
98
101
87
146
117
1 10
132
134
M7
106
162
168
135
183
94
118
114
|4|
152
132
123
Anaheim - Southern Zone Laboratory
(15 feet above ground)
Single Cascade Impactor
(20 CFM)
No. of
Sampl es
2
3
2
3
2
3
5
3
6
5
5
5
5
5
4
5
5
5
6
4
5
4
5
5
% of Mass !
< 1 .1 (j,m
27.1
39.0
32.5
37.0
49.9
42.3
37.7
49.1
39.5
23.5
22.3
27.5
23.3
28.6
26.2
25.5
26.4
25.2
28.7
37.6
28.0
36.9
26.5
28.8
27.9
23.7
25.0
28.2
30.4
% of Mass
< 3.3 nm
48.4
58.9
56.0
56.8
65.9
56.6
56.1
64.7
55.2
40.0
48.1
49.6
46.4
48.1
48.2
49.6
52.1
47.4
49.9
58.1
50.5
57.0
49.3
49.2
52.3
47.7
48.5
50.4
52.0
MMEDa)
(fim)
3-7
2.1
2.4
2.4
0.9
4.1
2.3
1.0
2.2
3.1
3.1
3.3
3.6
2.5
3.6
3.5
3-0
3-7
2.9
2.0
3.0
2.1
2.7
3.2
2.8
3.5
3-3
2.9
2.7
TSP
ftg/m'
l4i
142
176
134
133
142
i<*3
124
133
82
73
117
118
101
123
137
164
120
155
192
146
183
94
84
97
1 10
106
1 1 1
124
Single
Hi-Vol
Sampler
(50 CFM)
TSP ng/m'
85
103
122
101
117
128
111
86
88
77
55
102
100
83
101
119
146
104
136
179
127
158
70
76
84
93
103
97
102
a) Mass Median Effective Diameter,
b) No samples taken in August due to equipment repairs.
-------
Table II
SUSPENDED PARTICULATE MATTER ON EACH STAGE OF THE ANDERSEN CASCADE IMPACTORS
EXPRESSED AS PERCENTAGE OF TOTAL MASS COLLECTED —1975 CALENDAR YEAR
Month
(1975)
Jan.
Feb.
Mar.
Apr.
May
June
6 Mo. GM
July
Aug.
Sep.
Oct.
Nov.
Dec.
6 Mo. GH
Annual Geo.
Mean (AGM)
Los Angeles - Metropolitan Zone Laboratory
Sixth Floor Roof
(85 feet above ground)
Stage 1
> 7 tlm
18.6
2k. 8
17.1
29.6
22.7
26.0
22.2
24.9
23-7
16.7
34.2
27.0
23.2
24.4
23.5
Stage 2
3.3-7Mm
II. 0
14.1
8.0
lit. 6
12.5
15.8
12.1
13.4
11.9
11.7
15.0
12.8
12.2
12.8
12.6
Stage 3
2-3.3A>m
5.9
8.9
6.5
10.7
9.5
9.7
8.1
9.8
9.1
1 1 .1
7.3
9.3
8.8
9.2
8.7
Stage k
1 . l-2/m
7.2
6.6
6.5
8.2
9.9
12.4
8.2
10.3
12.9
12.5
8.9
8.8
9.7
10.4
9.2
Backup
7 M"1
27.8
33.0
34.7
25.8
24.1
30.8
29.5
26.4
30.0
28.7
35-9
29.5
26.1
29.3
29.4
Stage 2
3.3-7»m
14.6
16.7
15.5
11.4
11.8
15-3
14.2
13.6
15.2
17-3
14.1
13-5
12. 1
13.6
13.8
Stage 3
2-3. 3tm
5.0
1 1.6
10.4
9.2
8.0
9.2
8.6
9.3
10.4
8.8
8.9
7.6
9.2
9.0
8.8
Stage k
1 . l-2pn
9.2
9.4
9.5
6.9
8.0
1 1 . 1
9.0
9.7
11.6
10.1
8.4
8.4
7.3
9.1
9.0
Backup
7 (im
23.0
28.7
34.4
36.0
34.0
33-5
31.5
34.6
34.2
32.0
36.2
32.8
26.9
32.7
31.9
Stage 2
3.3-7JHH
10.5
13.1
14.8
13.1
15.4
19.2
14.2
16.2
|6.0
14.8
15.9
14.2
12.7
14.9
14.5
Stage 3
2-3. 3>m
8.8
8.5
10.3
14.7
11.9
11.4
10.9
11.7
12.3
12.5
11.4
11.2
10.5
11.6
11.2
Stage 4
1 . l-2fim
6.8
7.2
6.2
1 I.I
10.2
11.7
8.6
10.3
11.8
13.2
10.8
10.0
10.0
1C. 9
9.2
Backup
,
-------
LEAST SQUARES REGRESSION SUMMARY FOR
SINGLE YEARS AND FOR TWO-YEAR PERIOD
COMPARING IMPACTORS AND HI-VOLS
Type of Filter
for Back-up
and Hi-Vols
Period
Gel man A, 7/7/1-6/75
Spectro Grade
Gel man A, 7/7^-6/75
Spectro Grade
Location
Equation
No. of
Corre I. Samp I ing
Coeff. Pairs
Anaheim IMP = I.05HV +23.80 0.77
Los Angeles IMP
(85 feet)
= I.08HV + 0.002 0.91 60
Spectro Grade 7/75-6/76
Gel man AE
Anaheim IMP = I.03HV + 12.53 0.98
58
7/75-6/76 Los Angeles IMP = 0.76HV +27.68 0.96 56
(85 feet)
AlI of above
filters
All of above
fiIters
7/7^-6/76 Anaheim IMP = 0.97HV +25.^9 0.88 I
7/7^.6/76
Los Angeles
(85 feet)
IMP = 0.85HV + 20.^+5 0.91
02
116
In the linear regression summary table above, the reference "Gelman A,
Spectro Grade," means that both back-up and hi-vols at a particular loca-
tion had the same filter type, either Gelman A or Spectro Grade (see under
Experimental Procedures). In other words, during part of the period from
7/7^ to 6/75, Gelman A was used, and for the remainder of the period Spectro
Grade was used. At the time of changes in the type of filter materials, the
magnitude of effects upon the total particulate loadings wasn't known. For
the second year (7/75 - 6/76) it should be noted that correlation coeffi-
cients were excellent. The difference in the offset constants (12.53 vs.
27.68) could be due to the different filter types used at the two locations.
More recently, it has been shown by this District that TSP values, when
obtained in Los Angeles, bear the following relationship shown below for
different types of filter'. The TSP value obtained with Gelman AE filters
was arbitrarily given the relative ranking of 1.00. The diagram provides
a relative ranking for four filter types. Since all four types were not
compared directly, the percentage differences were derived by inference
based on the comparisons that were made.
VARIABILITY IN TSP LOADING WITH DIFFERENT HI-VOL FILTER MEDIA
20%
I
Spectro
Grade
(1.07)
7%
>
AE
(1 .00)
1%
>
EPA
(0.99)
12%
>
I
A
(0.89)
-9-
-------
The EPA filter-type* shown above was a special batch (No. 8788) of Gelman A
furnished to us by courtesy of the EPA. It was designed to contain mini-
mum trace metals and to yield a minimum of "artifact sulfate." The overall
value of 20% seen at the top refers to the difference between extremes and
is based on Gelman A as a reference (Gelman AE filters are currently used
in the Metropolitan zone and Spectro Grade in the Southern zone).
In the least squares computations summarized above, the samples for Ink to
6/75 at Anaheim exhibited a lower correlation coefficient for the relation-
ships between impactor (IMP) and hi vols (HV). Also, the regression equa-
tion had a large intercept (23.80) and that for Los Angeles had virtually
none (0.002). The difference is believed to be due to the fact that Gelman
A and Spectro Grade generally exhibit a 20% difference in the amount of
material captured as shown in the above comparison diagram taken from Re-
ference 9. Furthermore, the use of two types of filters for differing
periods at the two locations produces differing correlations, possibly
because one forms more artifact particulates than another. Until several
years of additional data are obtained, we must bear with the changes in
the filter materals which were varied more than desired during the first
year of the program. This must be borne in mind when drawing conclusions
even for TSP. When dealing strictly with sulfates, nitrates, or other chemi-
cal species, which are known to be formed in part by the artifact process,
the type of filter used in the future may be even more of a serious problem.
Effective July I, 1977, therefore, all hi-vol and back-up filters will be
Gelman A at all locations in the South Coast Air Quality Management District,
Fractionation Results
One simple way to show fractionation results is to abstract the 1975 geo-
metric mean data from Table II as shown below. This process puts the data
on a calendar year basis for convenience and possible future comparisons,
since this is a continuing survey. These data show that stage 3 has the
smallest percentage (hence mass) at both downtown Los Angeles locations.
This situation differs from Anaheim, where Stage k exhibited the least
mass. The total material captured (TSP) was nearly identical at Anaheim
and at the 22-foot level in Los Angeles (123 vs 122 jig/m3) for the year as
can be seen from the last column below.
ANDERSEN SAMPLER FRACTIONATIONS
(1975 Annual Geometric Mean Data)
No. of TSP
Location Samples Percent of Total Weight Collected
Stage I Stage 2 Stage 3 Stage 4 Back-up
. . . > 7 urn 3.3-7um 2. 0^3. 3am I .l-2.0um
-------
Raw data from an Andersen sampler generally are not sufficient for a study
of bimodality because the fractions are not of equal segments and there
are an insufficient number of fractions. We can conclude, however, that
the 85-foot level exhibits a larger preponderance (4l .5 percent) of the
-------
the minimum between modes is not apparent because there are too few frac-
tions. In Figure k the 1975 annual means are plotted as histograms (with
smooth curves connecting the segments to indicate trends only, with no
mathematical implication). Bimodality may be clearly ascertained for the
85-foot location with the minimum between modes at the I.I to 2.0 ym frac-
tion. This is probably due to the distorted particulate size population
at the higher elevation (larger particles are not uniformly distributed).
The minimum between the modes may exist at about I (jm at the other two loca-
tions, but bimodality could not be demonstrated. Notice that the Lundgren-
type mass plot10 is employed in Figure k where logarithmic artifacts are
used to accentuate the peak for the larger particle mode (above 7 ^m) which
otherwise would exhibit a long, drawn-out curve. In addition, dividing
the Amass by the log function, as done for the ordinate, accentuates peak
heights.
Four groups(refs. I 0,M, 1 2, 13) have suggested that the bimodal distribution for
atmospheric aerosols is a general rule. In the California ACHEX program the
minimum for bimodality at Los Angeles and Rubidoux was found to be 0.8 jum on
a particle-count basis12. Others have shown that the minimum in the bimodal
distribution ranges from one to two £fm on the basis of aerosol size-number
distribution13. The latter authors have reported that Andersen impactors
and the Minnesota Aerosol Sampler yield approximately the same results.
Total Suspended Particulate Air Quality Standards^
TSP as measured by hi-vol sampler is the federal reference method for de-
termining compliance with the air quality standards3. There are no State
or federal standards based on fractionation data. The Andersen impactor
TSP results differ from those by regular hi-vo! and would not be permiss-
ible for determining compliance currently, although a new promulgation of
such standards might be the next action to be expected in the area of TSP.
It is expected, however, to entail dichotomous sampling, not the multiple
cuts of the Andersen method14.
When one investigates the bases for certain standards, it becomes obvious
the aerodynamic separators have not been in prevalent use; hence there were
insufficient data on which to base size-related standards. The U. S. Public
Health Services criteria document on particulates discloses that health
hazards exist where the geometric annual mean TSP exceeds 80 ^/g/m3. Adverse
effects on materials (corrosion of steel and zinc) occurs at 60 ng/m3 15.
The following air quality standards for TSP are presently in effect:
Cal i form' a ARB
AIR QUALITY STANDARDS
(All values in uq/m3
< 60 annual geom. mean
< I 00 in any 2^-hour period
United States EPA
Primary: < 75 annual geom. mean
< 260 2^-hour
Secondary: < 60
< 150 2^-hour
-12-
-------
The hi-vol data reported in this paper showed the following annual geometric
mefans for 1975, (which was the only complete calendar year obtained).
TSP AS DETERMINED BY.HI-VOL
Annual Geom. Mean, uq/m3
Anaheim Los Angeles
(85 feet)
103 102
It is apparent that neither the California nor the federal annual standards
were met.
-13-
-------
19
9
e
y
«
s
STAGE 4
©
o
o
a
1
0.9
0.*
0.7
0.6
0.3
STAGE 3
For
Composite of Stages
and Back-up
STAGE 3
STAGE 1
BACK-UP
JUfME 1975
AM&HEIM
9* 95 90 SO 70 60 50 40 30 20 10
Percent of Mass Greater Than Size Indicated
Figure I. Typical log-normal plot of monthly geometric mean data from the Anaheim
station illustrates how the overall mass median effective diameters (MMED's) were
obtained for the monthly geometric mean Andersen impactor compilation in Table I.
-------
3.5
3.0
a.o
c
0
I IJ
1.0
o.s
•5 Feet Los Angelet
• — 22 Feet Los Angeles
--• 13 Feet Anaheim
JULY-SEP OCT-DEC JAN-MAR APR-JUN ^Oi-StP OCT-DEC JAN-MAR APR-JUN
1974 1975 1976
Quarter and Y«ar
Figure 2. Geometric means for quarterly periods show the difference in mass median
effective diameter (MMED's) between the stations.
-------
•
e
o
e
e
i
w
.«.
<*•
i
o
160
146
120
190
60
20
85 feet Lot Angela*
22 Feet Let Angeles
IS Feet Anaheim
S O
1974
N D
M
M
J J A
197S
Month and Year
MAM
1976
Figure 3. Monthly geometric mean mass loadings for Andersen impactor samples show a
general tracking of participate highs and lows for the three stations with no clear
seasonal patterns.
-------
< 30
>>.
s
<
10
\
Lot Angeles
85 Foot Location
1975 Geometric
Mean Data
Dp, Particle Aerodynamic Diameter (pm)
Lot Angeles
22 Foot Location
1975 Geometric
Mean Data
Dp, Particle Aerodynamic Diameter ( pm)
Anaheim
15 Foot Location
1975 Geometric Mean Data
1.1 2.O 3.3 ' TO
Dp t Particle Aerodynamic Diameter
Figure k. The mass-size distribution of total suspen-
ded particulates for downtown Los Angeles and Anaheim
are plotted using the mass-density function on the or-
dinate (note that bimodality can be ascertained only
for the 85-foot location).
-17-
-------
ACKNOWLEDGEMENTS
We thank Frank Hatanaka for diligence in performing the
collection and weighing of all samples in Los Angeles,
and Marjory Smith, Carl Anderson, Roger Griffin, and
Winston t-1isa for performing these tasks in Anaheim.
-18-
-------
REFERENCES
I. Wadley, M. W., and R. D. MacPhee, "Suspended Particulate Matter in the
Atmosphere of Los Angeles County, 1965 - 1974", Air Quality Report No.
82, Southern California Air Pollution Control District (SCAPCD) Metro-
politan Zone, May 1976.
2. Bope, W. G., unpublished report SCAQMD, Southern Zone.
3. Reference Method For The Determination Of Suspended Particulates In
The Atmosphere, Code of Federal Regulations, Part 40, P- 12 (July 1976).
4. Burton, Robert M., James M. Howard, Robert L. Penley, Peggy A. Ramsay,
and Thomas. A. Clark, "Field Evaluation of the High-Volume Particle
Fractionating Cascade Impactor," J. Air Pollution Control Assoc.,
23, 277-81 (1973).
5. Lee, Robert E., Jr., and Stephen Goranson, "National Air Surveillance
Cascade Impactor Network. I. Size Distribution Measurements of Suspen-
ded Particulate Matter in Air, "Environ. Sci. Technol.. 6_, 10] 9-1 024
(1972).
6. Rote, Donald M., and Lawrence E. Wangen, "Analysis of Suspended Parti-
culates in the Chicago Area by Particle Size," paper presented at the
68th Annual Meeting of the Air Pollution Control Association, Boston,
Mass., June 1975.
7- Hatch, T. F., and P. Gross, Pulmonary Deposition and Retention of Inhaled
Aerosols, Academic Press, New York, N. Y., 1964.
8. Wadley, M. W., J. A. Wood, and R. D. MacPhee, "Dependence of Hi-Vol Mea-
sur.ements on Airflow Rate," Southern California Air Pollution Control
District (SCAPCD), Metropolitan Zone, September 1975.
9. Witz, S., and R. D. MacPhee, "Effect of Different Types of Glass Filters
on Total Suspended Particulates and Their Chemical Composition^' J. Air
Pol Iution Control Assoc.. 2} J977 (in press).
10. Lundgren, Dale A., and Harold J. Paulos, "The Mass Distribution of
Large Atmospheric Particles," J. Air Pollution Control Assoc., 25,
1227-1231 (1975). ' ~~ "— —
II. Lee, Robert E., Jr., and Stephen Goranson,"National Air Surveillance
Cascade Impactor Network. III. 'Variations in Size of Airborne Parti-
culate Matter Over Three-Year Period," Environ. Sci. Technol., 10,
?022-1027 (1976). —
-19-
-------
12, Hidy, G. M., et. al., "Summary of the California Aerosol Characteriza-
tion Experiment." J. Air Pollution Control Assoc.. 25 1106-11\k (1975).
13. Willeke, Klaus, and Kenneth T. Whitby, "Atmospheric Aerosol Size Distri-
bution," J. Air Pollution Control Assoc. f 2£, 529-53** (1975).
\k, "Air duality Criteria For Atmospheric Lead" external review draft Nov.
19?6, U. S. Environmental Protection Agency, Washington, D. C.
15. "Air Quality Criteria For Participate Matter," U. S. Dept. of HEW,
Public Health Service, National Air Pollution Control Administration,
Washington, D. C., January, 1969.
-20-
------- |