EPA-600/2-77-132
July 1977
Environmental Protection Technology Series
GENERATION OF FUMES
SIMULATING PARTICIPATE
AIR POLLUTANTS
Industrial Environmental Research laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental Protection
Agency, have been grouped into five series. These five broad categories were established to
facilitate further development and application of environmental technology. Elimination of
traditional grouping was consciously planned to foster technology transfer and a maximum
interface in related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumenta-
tion, equipment, and methodology to repair or prevent environmental degradation from point
and non-point sources of pollution. This work provides the new or improved technology
required for the control and treatment of pollution sources to meet environmental quality
standards.
EPA REVIEW NOTICE
This report has besn reviewed by the U.S. Environmental Protection Agency, and approved
tor publication. Approval does not signify that the contents necessarily reflect the views and
policy of the Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
This document is available to the public through the National Technical Information Service,
Springfield, Virginia 22161.
-------
EPA-600/2-77-132
July 1977
GENERATION OF FUMES SIMULATING
PARTICULATE AIR POLLUTANTS
by
J.W. Carroz, F.K. Odencrante, and W.G. Finnegan
Research Department
Naval Weapons Center
China Lake, California 93555
EPA Interagency Agreement IAG-D5-0669
Program Element No. 1AB012
ROAP No. 21ADM-031
EPA Project Officer: Dennis C. Drehmel
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
DISCLAIMER
This report has been reviewed by the Industrial & Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for use.
ii
-------
FOREWORD
This project was undertaken to develop techniques for generation of
large quantities of reproducible, stable, inorganic, fine solid particle
aerosol fumes which are required for testing industrial pollution control
apparatus. Aerosols were generated to simulate the effluents from combus-
tion of pulverized coal, zinc smelters, and arc and basic oxygen furnaces.
This work was performed at Naval Weapons Center and funded by the
Environmental Protection Agency under Interagency Agreement EPA-1AG-D5-0669.
iii
-------
ABSTRACT
Thiis project supported the EPA stationary industrial pollution control
program. Techniques were developed for generating large quantities of repro-
ducible, stable, inorganic, solid particulate, fine particle aerosol fumes.
The aerosols were generated by burning flammable solutions containing appro-
priate soluble compounds (nitrates, for example) of the desired elements. In
the flamo these compounds decomposed to oxides. Particle size determinations
were made using scanning and transmission electron microscope (SEM and TEM)
photographic analysis of captured particles and Whitby and Royco aerosol
analyzers. Particle sizes from 0.0075 to 10 micrometers and concentrations up
to 1010 particles per cm3 were measured. Particle compositions were determined
by X-ray diffraction, SEM X-ray non-dispersive and cyclotron excitation analyses.
A dynamic aerosol diluter was designed to reduce the aerosol particle concentra-
tions by a factor as large as 38,700 so that the Whitby and Royco aerosol
analyzers would not become saturated. The generated aerosol flow rates were as
high as 42 m3 per minute (148 cfm); the particle loadings were as high as
16.8 g per m3 @ STP. For most aerosols the aerosol particle and condensation
nuclei concentrations were of the order of 109 particles per cm3. The aerosol
volume median diameters varied from less than 0.015 to greater than 4.7 ym and
were primarily a function of the solution ingredients.
The particle counts of the Whitby and Royco were compared. In the overlap
range of 'ihese instruments, the ratio of the Whitby counts to those of the Royco
varied from 0.68 to 343! For a coal fly ash simulation aerosol, the particle
counts of the EAA and Royco were compared with cascade impactor (Lundgren and
Celesco) neasurements.
Particle loadings (g/m3) were determined from the instrument particle
counts anc also from the generation rates. A comparison of the loadings shows
that for some of the generated aerosols the loadings measured by the Whitby were
significantly higher than the loadings determined from the generation rates.
Electron microscopic pictures (SEM and TEM) of precipitated particles show
that many of the larger particles (20 ym) are hollow and that the smaller
particles (0.01 ym) are in chain aggregates.
Special aerosols were generated to simulate the fine particulate effluents
generated by combustion of pulverized coal (electricity generation) electric
arc and basic oxygen furnaces (iron and steel production) and zinc smelters.
Methods were developed to vary the sulfur dioxide concentration and the particle
resistivities. The generation technique can be used to generate aerosols of
many different oxides and chlorides.
This work was funded by Interagency Agreement EPA-1AG-D5-0669 between NWC
and the EPAo It began in January 1975 and ended May 1977.
iv
-------
CONTENTS
Foreword ill
Abstract iv
Figures vi
Tables ix
Acknowledgment xi
1. Introduction 1
2. Conclusions and Recommendations 3
3. Characteristics of Effluents 4
Coal Fly Ash 4
Other Effluents 8
4. Method of Generation 11
Hardware 11
Solutions for Simulation Aerosols 15
Safety 15
5. Aerosol Sampling System 16
Generator 16
Diluter 17
Instruments 20
Size Distributions 22
6. Results 23
The Particles 23
Particle Counts 25
Volume Distributions . 38
Particle Compositions 41
Coal Fly Ash Fume Simulation Aerosol 43
Electric Arc Furnace Fume Simulation Aerosols 55
Basic Oxygen Furnace Fume Simulation Aerosols 59
Zinc Smelter Fume Simulation Aerosol 62
7. Costs of Raw Materials 65
8. Discussion of Results 67
9. References 74
Appendices
A. Operating Specifics 76
B. Aerosol Size Distributions 79
C. Particle Concentrations and Volume Percents 83
-------
FIGURES
Number
1 Airborne fly ash volume distribution of the submicron particles
v.pstream of the baghouse at the Nucla Power Plant
2 Generator set-up for coal fly ash simulation aerosol. The larger
diameter stack reduces the exit aerosol temperature 12
3 Generator set-up for producing simulation aerosols. Reducer is
sometimes used at top of stack to increase solids loadings. . . 13
4 Three stage dilution set-up. Each stage dilutes an aerosol a
fixed amount. Dilution is varied by using 1, 2, or 3 stages
in series 18
5 Diluter, Stage #1 19
6 Typical background measurements before aerosol generation. The
background measurements shown in this figure are multiplied by
6900 because with three stage dilution the equivalent of 6900
parts of air is mixed with one part of aerosol 27
7 Comparison of measurements with the electrical aerosol analyzer
(EAA) of two SiOa aerosols, one containing 0.163 cm3/m3 of
aerosol, another containing 0.004 cm3/m3. The EAA does not
appear to have been saturated 32
8 Comparison of measurements with the Royco optical particle counter
and the EAA of the 0.78 g/m3 FeaOa & Fe30n aerosol. The measure-
ments with the Royco & EAA are in close agreement 37
9 Comparison of measurements with the Royco and EAA of the 0.054
g/m3 MgO aerosol. The measurements with the Royco and EAA do
not agree with each other 37
10 The volume distribution determined from EAA and Royco measure-
ments of the 0.62 g/m3 MhaOi, aerosol. Most of the volume of
aerosol particles appears to be included in the size range
measured by the instruments 39
11 The volume distribution determined from EAA and Royco measure-
ments of the 0.69 g/m3 CrzOa aerosol. A significant portion
of the volume of aerosol particles probably exists as particles
larger than 10 ym 39
vi
-------
FIGURES (continued)
Numbers Page
12a The volume distribution determined from EAA & Royco measure-
ments of the 1.8 g/m3 FeaOij aerosol. This aerosol was
found to have a volume mode of larger (D>10 urn) particles. . . 40
12b FeaOit aerosol 1.8 g/m3, the volume distribution of some large
particles captured on a glass slide. Possibly 80-90% of
the aerosol mass existed as large hollow particles ...... 40
13 The volume distribution determined from EAA & Royco measure-
ments of the 0.41 g/m3 A1203 2HaO, FesOi, aerosol. This
aerosol has a bimodal distribution 41
14 The size-distribution, plotted as AN/AlogD vs. D, from measure-
ments with the EAA and Royco of the coal fly ash simulation
aerosol generated 22 November 1976 46
15 The size distribution from measurements with the EAA and Royco
of the fly ash simulation aerosol generated 22 November 1976 . 47
16 The size distribution from measurements with the EAA and Royco
of the fly ash simulation aerosol generated 3 December 1976. . 47
17 The volume distribution determined from the EAA and Royco
measurements of the fly ash simulation aerosol generated
22 November 1976 , 48
18 The volume distribution determined from the EAA and Royco
measurements of the fly ash simulation aerosol generated
3 December 1976 48
19 The mass distribution of the fly ash simulation aerosol measured
with a Celesco cascade impactor assuming a particle density
of 2 g/m3. (Courtesy of Dr. R. L. Chuan, California Measure-
ments Inc. Sierre Madre, CA. Note how this mass distribution
compares with the volume distribution shown in Figs. 17 & 18 . 50
20 The coal fly ash simulation aerosol size distribution as a
function of particle size as determined with a cascade
impactor (mass distribution) and with the EAA and Royco
(volume distribution) 51
21 The size distribution of the coal fly simulation aerosol with
Li2COs added to increase the particle conductivity. The size
distribution is similar to that of the simulation aerosol;
adding the Li2C03 had a small effect 53
22 The volume distribution determined from the EAA and Royco
measurements of the coal fly ash simulation aerosol with
Li2C03 added 53
vii
-------
FIGURES (Continued)
Numbers Page
23 Ei.ectric arc furnace simulation aerosol A (0.65 g/m3) size
distribution from measurements with the EM and Royco .... 57
24 Electric arc furnace simulation aerosol B (2.1 g/m3) size
distribution from measurements with the EAA and Royco .... 57
25 Tt.e volume distribution determined from EAA and Royco measure-
ments of electric arc furnace simulation aerosol A
(0.65 g/m3) 58
26 The volume distribution determined from EAA and Royco measure-
ments of electric arc furnace simulation aerosol B
(1.2 g/m3) 58
27 Electric arc furnace simulation aerosol A (0.65 g/m3) size
distribution, plotted as AN/AlogD vs. D 59
28 Ths measured size distributions of the electric arc furnace
simulation aerosols A and B are nearly the same 59
29 Basic oxygen furnace simulation aerosol AA (Fe20a 6 g/m3)
isize distribution plotted as AN/AlogD vs. D 61
30 Basic oxygen furnace simulation aerosol AA (FeaOs 6 g/m3)
jsize distribution from measurements with the EAA & Royco. . . 61
31 The volume distribution determined from EAA and Royco measure-
rients of the basic oxygen furnace simulation aerosol AA.
Ninety-five percent of the particle volume was contained in
particles with diameters between 0.15 and 3 ym „ . 62
32 ZnO (0.94 g/m3) aerosol size distribution from measurements
vi'ith the EAA and Royco . 63
33 ZnO (0.94 g/m3) aerosol size distribution plotted as AN/AlogD . 63
34 The volume distribution determined from EAA and Royco measure-
ments of the ZnO (0.94 g/m3) aerosol 64
viii
-------
TABLES
Number Page
1 Pulverized Coal Combustion Effluent (Electricity Generation) . . 5
2 Elemental Composition 7
3 Electric Arc Furnace Effluent (No Oxygen Lance). . 8
4 Basic Oxygen Furnace Effluent. 9
5 Zinc Roaster & Sintering Machine Conditions at Precipitation . . 10
6 Zinc Roaster & Sinter Machine Effluent 10
7 Diluter Calibration, Stage 2 17
8 Aerosol Particle Loading (Weight or Volume/Hot Gas Volume)
Measured in 0.425 m Stack with Three Stage Dilution 30
9 Aerosol Particle Loading (Weight or Volume/Hot Gas Volume)
Measured in 0.25 m Stack with Three Stage Dilution 31
10 Aerosol Particle Concentration (CNC vs EAA) Measured in 0.25 m
Stack Using Three Stage Dilution 34
11 Aerosol Particle Counts. Measured in 0.42 m Stack with Three
Stage Dilution . 35
12 Aerosol Particle Counts. Measured in 0.25 m Stack with Three
Stage Dilution 36
13 Particle Compositions 42
14 Characteristics of the Coal Fly Ash Simulation Aerosol 43
15 Cyclotron Excitation Analysis of Lundgren Impactor and Nuclepore
Filter Samples of the Coal Fly Ash Simulation Aerosol 44
16 Measured Mass Distribution of the Coal Fly Ash Simulation
Aerosol (0.66 g/m3 of Aerosol is Generated) 49
17 Impactor Data (Diameter Range Based on a Density of 2 g/m ). . . 50
18 Particle Resistivity* of Coal Fly Ash Simulation Aerosol
(T=21°C, RH=27%) 52
ix
-------
TABLES (continued)
Numbers
19 Characteristics of Electric Arc Furnace Simulation Aerosol A . . 55
20 Characteristics of Electric Arc Furnace Simulation Aerosol B . . 56
21 Characteristics of Basic Oxygen Furnace Fume Simulation
Aerosol AA 60
22 Characteristics of Basic Oxygen Furnace Fume Simulation
Aerosol BB 60
23 Characteristics of the Zinc Smelter Fume Simulation Aerosol. . . 63
24 Estimated cost of ingredients to generate 45 kg (100 pounds)
aerosol particles to simulate the fine particulate effluent
from the following industrial sources 66
25 Optical Properties 69
26 Comparison of Coal Fly Ash Effluents with Simulant . » 71
27 Comparison of Electric Arc Furnace Effluents with Simulants. . . 72
28 Comparison of Basic Oxygen Furnace Effluents with Simulants. . . 72
29 Comparison of Zinc Roaster and Sintering Machine Effluents with
Simulant 73
-------
ACKNOWLEDGMENT
The cascade impactor measurements by Dr. Raymond L. Chuan; constructive
criticism of a preliminary diluter design by Professor Kenneth T. Whitby;
advice on the simulation of industrial effluents by Drs. Dennis C. Drehmel
and James H. Abbott; guidance on particle measurements by Dr. Edward E.
Hindman II, and the leadership of Dr. Pierre St.-Amand are gratefully
acknowledged.
xi
-------
SECTION 1
INTRODUCTION
Particles smaller than 5 ym (fine particles) can penetrate the human nasal
cavity and enter the lungs. Fine particles remain airborne for long periods
because they are too small to rapidly settle to the ground. Thus, the
potential hazard of these particles as an air pollutant is greater than that
of larger particles.
Pollution abatement equipment is available for collecting large particles
(D>5y). Conventional technology can be used to control primary fine particle
emissions (Drehmel, 1976a). When the resistivity of the particles is about
1010 ohm-cm, some electrostatic precipitators collect 90% of the particles at
all size ranges. The collection efficiency of electrostatic precipitators,
however, depends on the particle electrical resistivity. Since electrostatic
precipitators are the most common particulate control devices in power
production and since western low sulfur coals produce a significantly higher
resistivity fly ash, the low sulfur coals will make it more difficult to
conform to particulate emission standards for utility boilers with standard
equipment (Drehmel, (1976b). Fabric filters can collect 95% of the particles
at all size ranges and do not depend on particle resistivity. There are
possibilities for improvements in all precipitators and filters. Emerging
technology should improve the collection and reduce operating costs.
The Environmental Protection Agency needs stable sources of well-
characterized solid particulate, fine particle aerosols (fumes) for laboratory
pilot plant tests of pollution abatement equipment. To provide this, we have
conducted a research program on the continuous generation of inorganic solid,
fine particulate aerosols (fumes). These aerosols are intended to simulate
the fine particulate effluents generated by combustion of pulverized coal for
electricity generation, electric arc and basic oxygen furnaces, and zinc
smelters. Only the fine particle (D<3ym) portion of the effluents were to be
simulated. For coal fly ash simulation aerosols, the particle electrical
resistivity was to be variable over a wide range.
The aerosols were generated by burning flammable solutions containing
appropriate soluble compounds (nitrates, for example) of the desired elements.
In the flame these compounds decomposed to oxides. Particle size determinations
were made using scanning and transmission electron microscope (SEM and TEM)
photographic analysis of captured particles and Whitby and Royco aerosol
analyzers. Particle sizes from 0.0075 to 10 micrometers and concentrations
up to 1016 particles per cubic meter could be measured. Particle compositions
were determined by X-ray diffraction, SEM X-ray non-dispersive and cyclotron
excitation analyses.
-------
This work was funded by Interagency Agreement EPA-1AG-D5-0669 between
NWC and the EPA. It began in January 1975 and ended April 1977.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
The program objective of developing reproducible stable sources of well-
characterized solid particulate, fine particle aerosols for laboratory pilot
plant tests of pollution abatement equipment was accomplished. The materials
and the techniques used in this simulation aerosol project are flexible; many
thousand different aerosols can be made by choosing ingredients, dissolving
them in a flammable solvent, and burning them in the aerosol generation
apparatus. It will be possible to generate aerosols with a wide variety of
properties to meet specific test requirements.
The most useful aerosols are those comprised of a single compound because
all of the particles have many of the same properties (density, resistivity,
etc.) and this will simplify test analyses. By using several different
simple aerosols in separate tests, it should be possible to discover and
determine the basic principles and parameters concerning pollution abatement
equipment. Because of analytical difficulties, using complex aerosols comprised
of a variety of chemical compounds will not yield conclusive results.
The analysis of the results showed that for some aerosols the particle
mass loadings determined from the Whitby Aerosol Analyzer (EAA) were signif-
icantly higher than the particle loadings determined from the generation
rates. This finding is important if EAA counts are to be used to determine
aerosol particle mass loadings.
-------
SECTION 3
CHARACTERISTICS OF EFFLUENTS
COAL FLY ASH
The; characteristics of pulverized coal fly-ash and other coal fly-ash
vary widely. The type of coal, the plant and the pollution control devices
all affect the characteristics of the fly ash samples collected. Most analyses
to detemiine the characteristics of fly ash have been conducted with samples
of ash collected by pollution control devices. This material usually contains
coarse particles. There is a need for more analysis of samples containing
only the fine particles.
Tatle 1 (Vandegrift et^ a^., 1971) shows the characteristics of pulver-
ized coal fly-ash. Note the great variability in properties of different
ashes. The characteristics of the fine particles may not be the same as
those shown in Table 1 where 75% of the mass is larger than 5 ym.
Lute (1961) characterized a fly ash sample taken from a hopper of an
electrostatic precipitator. The ash contained 79% glass, 6% magnetite-
hematite and 4% carbon particles. His X-ray diffraction results showed
magnetite (FesOO the most abundant crystalline component, hematite (Fe203)
second, and quartz (Si02) third. Trace amounts of mullite (3A1203«2Si02),
calcite (CaC03) and anhydrite (CaSOO were found.
Harvey (1971) measured the fly ash from boiler number 10 of the TVA
plant in Paducah, Kentucky. He separated the ash into non-magnetic and
magnetic fractions. An X-ray diffraction analysis of the non-magnetic frac-
tion (48% of ash) was high in quartz, medium in lime and low in portlandite
(Ca(OH)a) and anhydrite. The analysis of the magnetic fraction was high in
magnetite, medium in hematite and low in periclase (MgO). The chemical
analysis of the fly ash showed 39% Si02, 16% A1203, 32% Fe203, 11% CaO, 2%
others.
Yakowitz et al. (1972) analyzed fly ash particles taken from the area of
an electric power plant in Minneapolis, Minnesota, by means of combined
electron microscopy and X-ray micro-analysis. For the most part, the fly ash
consisted of nearly spherical particles ranging in size from 0.5 to 3 ym or
more in diameter. The elements in the ash particles were silicon, aluminum,
calcium, iron, potassium, magnesium and titanium, listed in decreasing order
of abundance within the particles. The spherical particles did not vary
significantly in composition as a function of size.
-------
TABLE 1. PULVERIZED COAL COMBUSTION EFFLUENT (ELECTRICITY GENERATION),
reprinted with permission from Vandegrift, 1971, p. 53.
Particle size (wt.% < size, ym) - 15<3, 25<5, 42<10, 65<20, 8K40
Solids loading (g/m3) - 2.3 - 12.8 (mean 7.6)
Chemical composition (Wt.%) - Si02 17-64 (average 43)
Fe203 2-36 (15)
A1203 9-58 (24)
CaO 0.1-22 (4)
MgO 0.1-5 (1.0)
Na20 0.3-4 (0.9)
Bulk density - .48-.8 (g/cm3)
Particle density - 0.6-3.0 (average 2.3)
Electrical resistivity - 108-1013 ohm.cm
Moisture content - 0.23 wt.%
Carrier Gas
Temperature - 121-157°C
Moisture content - 5.9-8.0 Vol.%
Chemical composition (Vol.%) - C02
C02
02
N2
CO
NO
S02
S03
12.5-14.9
4.2-6.6
Balance
5-69 ppm
161-526 ppm
1080-1780
3-66ppm
ppm
-------
X-ray diffraction results of Cavin et al. (1974) revealed the mineral
content of the fly ash they measured to be primarily mullite with the presence
of hematite, magnetite, quartz and gypsum.
Davison et al. (1974) measured the elemental composition of several size
fractions of fly ash. Two types of fly ash are represented; (a) fly ash
retained in a cyctonic precipitation system, and (b) airborne fly ash col-
lected in the ducting about 3 m from the base of the stack with an Anderson
stack sampler. Table 2 shows the results of these measurements.
Sem (1975) measured the size distribution of airborne particles from 1.0
to less than 0.01 ym diameter at the Nucla, Colorado, power plant. The
results of Sem are shown in modified form in Figure 1 with his permission.
In Sem (1975), the volume distribution is reported as ym3/m3 since a consider-
able unknown fraction of the volume was above 1 ym. Figure 1 shows the
volume distribution of only the submicron particles upstream of the baghouse
as measured on November 12, 1975.
Shei et^ a±. (1977) characterized the emissions from a coal-fired power
plant with an electrostatic precipitator. They found about 35% by mass of
the particulates emitted were less than 3 ym (measured with an Anderson stack
cascade impactor). The major elements emitted were Fe, Ti, Si, S, K, Ca.
These emissions do not contain as much aluminum as the airborne fly ash
measured by Davison (see Table 2).
OU
54
48
42
36
UJ
| 30
O
> 24
18
12
6
0
-
-
-
-
r~
-
-
-
-
0.01 0.10 1.00 10.00
DIAMETER, inn
Figure 1. Airborne fly ash volume distribution of the submicron
particles upstream of the baghouse at the Nucla Power
Plant.
-------
TABLE 2. ELEMENTAL COMPOSITION, reprinted with permission from
Environmental Science & Technology, 1974 by Davison.
Copyright by the American Chemical Society.
Particle
>74
44-74
A. Fly Ash Retained in
Sieved fractions
diameter, ym Fe,wt.% Si,wt.% Mg,wt.2
18 18 0.39
Plant
1 C,wt.% S,wt.% Al.wt.%
• •• ••• *••
11 1.3 9.4
Aerodynamically sized fractions
>40
30-40
20-30
15-20
10-15
5-10
<5
>11.3
7.3-11.3
4.7-7.3
3.3-4.7
2.06-3.3
1.06-2.06
0.65-1.06
50 3.0 0.02
18 14 0.31
...
...
6.6 19 0.16
8.6 26 0.39
...
B. Airborne Fly Ash
13 34 0.89
...
12 27 0.95
...
17 35 1.4
...
15 23 0.19
0.12 <0.01 1.3
0.21 0.01 6.9
0.63
^ • 3 ••• •••
6.6 4.4 9.8
5.5 7.8 13
0.66 8.3 19.7
0.70
0.62 7.9 16.2
0.57
0.81 25 21.0
0. 61 ... ...
48.8 9.8
-------
OTHER EFFLUENTS
Table 3 (Vandegrift et^ al., 1971) shows that the characteristics of the
fumes f::om electric arc furnaces vary a great deal. The mass median diameters
of several electric arc furnace fumes have been measured (unknown author,
1974); i:hese diameters ranged from 0.3 to 5.4 ym.
Table 4 (Vandegrift et al., 1971) shows the characteristics of the fumes
from basic oxygen furnaces. The number median diameter of the fume is 0.012.
At a solids loading of 10 g/m3, there would be 3xl018 particles/m3 (using a
particle density of 3.44 g/cm3)!
Harris and Drehmel (1973) measured the solids loading of various sized
aerosol particles of zinc smelters using a Brink impactor. Table 5 shows
their results. Table 6 (Vandegrift et^ al., 1971) shows the characteristics
of the ::umes from zinc smelters.
TABLE 3. ELECTRIC ARC FURNACE EFFLUENT (NO OXYGEN LANCE),
reprinted with permission from Vandegrift, 1971,
p. 123 and 126.
Particle size (wt%
-------
TABLE 4. BASIC OXYGEN FURNACE EFFLUENT,
reprinted with permission from
Vandegrift, 1971, p. 123 & 126.
Particle size (wt %
-------
TABLE 5. ZINC ROASTER & SINTERING MACHINE CONDITIONS AT PRECIPITATOR,
reprinted with permission from Harris and Drehmel, 1973.
Particle Size Roaster Sintering Machine
ym
>3.1
1.8-3.1
1.25-1.8
0.62-1,25
0.38-0.62
<0.38
g/m3
Test 1
0
0.0032
0.0032
0.0708
0.1904
0.0551
g/m3
Test 2
0
0
0.0016
0.0142
0.0960
0.0730
g/m3
Test 3
0.1747
0.2817
0.6594
0.9820
0.6279
0.6720
TABLE 6. ZINC ROASTER & SINTER MACHINE EFFLUENT, reprinted with
permission from Vandegrift, 1971, p. 263-265.
Particle size (wt%
-------
SECTION 4
METHOD OF GENERATION
Solution-burning aerosol generators used for weather modification field
experiments (Carroz, 1972) have been modified to dispense metal oxides.
Flammable solutions of appropriate chemicals were made. The solutions were
put into a heavy walled, stainless steel tank, pressurized with nitrogen gas,
and, using standard industrial nozzles of the proper type and size, atomized
into aerosol droplets. The aerosol droplets were burned to produce pre-
dominantly submicron particles of metal oxides by high temperature decompo-
sition and/or vaporization of the chemicals. The hot gaseous products of
combustion, together with excess air, greatly reduce agglomeration of the
aerosol particles.
HARDWARE
Figures 2 and 3 and Photographs 1 and 2 show the aerosol generator and
stacks. The larger diameter stack shown in Figure 2 was used for pulverized
coal effluent simulation. The smaller diameter stack (Figure 3) was used for
the other simulations. For some aerosols, a reducer was used at the top of
the stack to decrease the draft and to increase the particle loadings. The
gas line pressurizes the pressure tank at 1.4 MPa (200 psi). Several sizes
of pressure tanks have been used, as small as 1 liter and as large as 38
liters. The tanks were fabricated from No. 316 stainless steel. The valve
located between the tank and the nozzle was used to turn the flow on and off.
The nozzle is an industrial solid cone nozzle with a 30 spray angle. The
flow rate of solution through the nozzle is about 5.9 x 10~6 cubic meters per
second (A Delavan Corp. Type WDB 4.0, 30 , stainless steel nozzle is satis-
factory). The flow rate depends on the solution being sprayed.
The flame cone (see Figure 2) consists of a perforated metal cone, 0.60
m high, with a 0.26 m diameter at the top and a 25 mm diameter at the bottom.
The cone is perforated with several 25 mm diameter holes. The flameholder
consists of a 0.25 m diameter, 0.50 m high sheet metal stove pipe.
The large diameter stack is 0.425 m in diameter and is 3.05 m high.
Ambient air is used to cool the combustion products coming out of the flame
holder. Jets of air were used to insure adequate mixing of the generated
aerosol fume and ambient air. There are four jets evenly spaced at the
bottom edge of the stack. The 1 mm inside diameter jets are about 45 from
the horizontal plane; they were supplied with air at 210 KPa (30 psi). The
small diameter stack is 0.25 m in diameter and 3.5 m high. It does not
require air jets. When the reducer was used, the diameter was reduced to
0.15 m.
11
-------
-------
FLAME HOLDER.
FLAME CONE'
.STACK
s
GAS LINE
PRESSURE TANK FOR
SOLUTION
Figure 3. Generator set-up for producing simulation
aerosols. Reducer is sometimes used at top of stack
to increase solids loading.
13
-------
Photograph 1.
Aerosol Generator and Stack for
Coal Fly Ash Simulation Aerosol.
Photograph 2.
Basic Oxygen Furnace Simulation
Aerosol Exiting from Stack Reducer.
-------
SOLUTIONS FOR SIMULATION AEROSOLS
The basic solution for generating simulated fly ash aerosol contains:
200 g of hydrated aluminum nitrate (A1(N03)«9H20); 91.7 g of hydrated ferric
nitrate (Fe(N03)3-9H20); and 112 g of dimethyl diethoxysilane ((CH3)2Si(OC2H5)2)
per liter of ethanol. Up to 80 g per liter of lithium nitrate (LiN03) can
be added to the basic solution to generate aerosol particles with a higher
electrical conductivity. Hydrated lithium perchlorate (LiC10i,'3H20) is also
compatible with the basic solution and was also used to raise the particle
conductivity. Carbon disulfide (CS2) can also be added to the basic solution.
Burning 30 ml of CS2 per liter of ethanol results in about 2300 PPM of S02 in
the aerosol.
When the solution for simulated fly ash is burned in air, the reactants
and reaction products are:
A1(N03)3'9H20 + Fe(N03)3-9H20 + (CH3)2Si(OC2H5)2 + C2H60 + (excess)02 ^
A1203-XH20 + Fe203 + Fe30i, + Si02 + C02 + H20 + N2
The basic solution for generating simulated electric arc furnace effluent
contains: 115 g of hydrated ferric nitrate (Fe(N03)3«9H20); 47.5 g of hydrated
chromic nitrate (Cr(N03)3*9H20); 43.5 g of hydrated magnesium nitrate (Mg
(N03)2«4H20); and 33 g of aqueous manganese nitrate solution (51% Mn(N03)2)
per liter of ethanol.
There are two basic solutions for generating simulated basic oxygen
furnace effluent. The first yields a high solids loading (6 g/m3) of hematite
particles (Fe203); the second yields a solids loading of 1.8 g/m3 of primarily
magnetite particles (Fe3OO. The first solution contains 293 ml (427g) of
iron pentacarbonyl (Fe(CO)5) in 700 ml of acetone (CH3COCH3). The second
contains 379 g of hydrated ferric nitrate (Fe(NOa)3«9H20) and 16.5 g of
hydrated calcium nitrate (Ca(N03)2*4H20) per liter of ethanol.
The basic solution for generating simulated zinc smelter effluent con-
tains 159 g of (73% pure) zinc nitrate (Zn(N03)2-xH20); 60 g of hydrated
cadmium nitrate (Cd(N03)2'4H20); 14.9 g vanadium oxide sulfate (VOSOil«xH20)
and 20 ml of water per liter of methanol.
An Operating Procedure in Appendix A gives some additional details about
the solutions.
SAFETY
Fe(CO)5 is considered very toxic (Brakhnova, 1975). CdO, CS2 and CH3OH
are also toxic; (CH3)2Si(OC2Hs)2 is probably toxic. Most heavy metals can be
toxic; when they exist as fine particle aerosols their potential hazard is
increased.
The Operating Procedure in Appendix A should be followed for safe
operation.
15
-------
SECTION 5
AEROSOL SAMPLING SYSTEM
GENERAL
The generated aerosol was sampled at the top of the stack. For particle
size measurements, the aerosol was diluted continuously at the top of the
stack, piped down to the laboratory (located just inside the building beside
the generator) and simultaneously measured with electrical and optical particle
counters and a condensation nuclei counter.
A water-cooled nickel plated 0.14 x 0.16 m rectangular brass plate was
used to precipitate aerosol particles for X-ray diffraction analysis. After
capture, a part of the sample was immediately sealed in a glass capillary
tube to await analysis. For SEM and TEM analysis the aerosol particles from
the top of the stack were collected on grids with a thermal impactor.
The fume velocity in the stack was measured with a specially designed
balance (see Photograph 3); the temperature was measured with a thermocouple.
Photograph 3.
Balance in position for measuring the fume velocity in the stack.
The stack is in the bottom right hand corner of the picture.
L6
-------
DILUTER
A relatively simple diluter system was used to reduce the aerosol particle
concentration so that electrical aerosol analyzers and optical particle
counters could operate without saturating. The diluter system dilutes an
aerosol in increments. The system has three diluter stages in series; one or
more stages can be used by connecting a sample line leading to the instruments
to the outlet of any of the three stages (see Figure 4 and Photograph 4).
The system was calibrated by generating several aerosols with different
particle concentrations. With the proper aerosol particle concentration in
the stack, measurements were made with the particle counters using x dilution
stages and using (x-1) stages. The aerosol particle concentration with (x-1)
diluter stages divided by the particle concentration with x stages is equal
to the aerosol dilution caused by stage x. The dilution was then computed
for many different particle sizes between 0.03 to 5 ym. Table 7 shows the
results of the measurements with 1 and 2 stages. The dilution by stage 1 is
16, by stage 2, 16 and by stage 3, 27. Therefore, if only the first stage is
used, the dilution is 16. If the first two stages are used, the dilution is
256 and if all three stages are used the dilution is 6900.
TABLE 7. DILUTER CALIBRATION, STAGE 2
Solution: 75g A1(N03)3'9H20 dissolved in 10 H of ethanol
Average Electrical Aerosol Analyzer Measurements,
Number of Particles with Diameters between D and D
x y_
Diameter (ym) 0.03 .06 .1 .15 .3 .6
ANi(#/cm3) 1 stage 1.6E4* 5.4E3 3.8E3 5.6E3 8.2E2
AN2(#/cm3) 2 stages
1.4E3 3.5E2 2.8E2
3.9E2 5.3E1
Ni/N2
11
15
14
14
15
Average Royco Measurements, Number of Particles
with Diameters Greater than D
x
Diameter
Ni(#/cm3)
N2(///cm3)
Ni/N2
(ym)
1
2
stage
stages
.8
95
6.6
14
1.
73
4.6
16
1.2
54
3.4
16
1.5
40
2.3
17
2.
28
1.6
17
3.
19
1.0
19
4.
11
0.56
18
5.
6.
0.
15
5
44
*Expodential notation, 1.6E4 is 1.6xlOl
17
-------
FUME
INTAKE
STAGE 1
r T
STAGE 3
STAGE 2 -^
Figure 4. Three stage dilution set-up.
Each stage dilutes an aerosol a fixed
amount. Dilution is varied by using 1,
2, or 3 stages in series.
DILUTED
FUME TO
INSTRUMENTS
Photograph 4. Aerosol dilution apparatus
L8
-------
Using particle counting instruments to measure the dilution is more
reliable than measuring the volumes of fume and air being mixed. Because of
the small pressure differences between stages, it is difficult to accurately
measure the gas flow rates without disturbing them. Accurately determining
the dilution of each stage is important because the total dilution is the
product of the dilution of each stage and any errors are multiplied.
Three different aerosols were used to measure the dilution of the three
stages. The stage 1 dilution was determined with an aerosol generated by
burning a solution of 0.05 g of zinc nitrate and 0.1 g of aluminum nitrate
per liter of ethanol. The stage 2 dilution was determined by burning a
solution of 7.5 g of aluminum nitrate per liter of ethanol and stage 3 by
burning a solution of 200 g of aluminum nitrate and 91.6 g of ferric nitrate
per liter of ethanol. The 0.42 m stack was used; the aerosol temperature at
the diluter intake was between 170 and 250°C for the various tests.
Concentrated aerosol enters the first dilution stage (see Figure 5)
through a 11 mm inside diameter stainless steel sample probe. The sampling
is close to isokinetic; the small differences in stack velocity and probe
velocity should not be significant for particles smaller than 10 ym diameter.
A streamlined reducer (reducer B shown in Figure 5) is attached to the end of
the tube to reduce the velocity through the tube. Reducers are used to
adjust the amount of dilution and to adjust the velocities for nearly iso-
kinetic sampling. Aerosol enters the diluter because the pressure at B is
less than at A. Ambient air enters the diluter at C and passes through a
sintered porous bronze tube D before coming into contact with the aerosol.
This uniform flow prevents any severe impaction or turbulence. Restrictions
at the inlet to C are used to adjust the pressure at B so that the aerosol
flows from one stage to the next. E is a mixer composed of two disks, each
with six holes. The disks are oriented to produce maximum turbulent mixing.
DILUTION AIR
INTAKE
ELBOW
i POROUS N BRASS END
METAL TUBE PIECE
Figure 5. Diluter, Stage //I.
19
-------
The coppar mixing pipe, F, extends from the downstream disk to the diluted
aerosol sampling point, G. A streamlined reducer (shown at G in Figure 5) is
attached to the copper tube located on the diluter centerline where the
diluted aerosol exits. The diluted aerosol (position H in Figure 5) is drawn
either into the next dilution stage or to the measuring instruments. The
remaining diluted aerosol leaves the diluter stage at I and is exhausted into
the atmosphere by a fan common to all stages. Filtered air from the labora-
tory is used for the aerosol dilution. There is no contamination of this air
by the generated aerosol. The air flow is adjusted by an orifice located
downstream of I. Stages 1, 2, and 3 have different sized orifices, designed
so that t:he vacuum from one fan maintains the proper pressures in all three
diluters. The feed line from the diluter to the instruments is connected at
H on stafte 3, or on stage 2 or 1 if less dilution is required. (See Appendix
A for a more detailed description of the diluter.)
INSTRUMENTS
The instruments received a continuous supply of fresh aerosol from the
feed line. The concentrated aerosol was rapidly diluted near the sample
point at the top of the stack. This diminished the rate of particle coagu-
lation. The diluted aerosol was transported to the laboratory through 6.86 m
of 13.3 urn I.D. copper tubing. The aerosol passed through a Thermo-Systerns
Model 3012 Krypton 85 aerosol neutralizer prior to entering the copper feed
line to the instruments. The flow rate through the copper tubing was about
22 liters per minute; half of this flow passed into the instruments. Short
lengths (~2cm) of plastic tubing connected the copper lines to the instruments,
thus essentially all the surfaces in contact with the aerosol were metal.
The only exception was aim long rubber tube which connected the condensation
nuclei counter to the sample line. Three instruments were used simultaneously
(see Photograph 5): an Environment I, Rich 100, condensation nuclei monitor
(CNC), a Model 3000 Whitby Electrical Aerosol Analyzer (EAA)*, a Model 200
Royco Optical Particle Counter (Royco). The CNC detects very small particles
(to about 20A in diameter) over a wide concentration range (from about 100
particles/cm3 to 107 particles/cm3). The EAA measures an equivalent sphere
diameter based on electrical charge and mobility. The Royco measures the
equivalent light scattering diameters of latex spheres. The Royco did not
have a sheath air inlet system to prevent aerosol recirculation in the sensing
zone. Thu Whitby-Cantrell Electrical Aerosol Analyzer Constants, adjusted to
the Model 3000, were used to obtain aerosol particle size distributions from
the currents measured with the EAA (see Appendix C). The original EAA Model
3000 data reduction constants are also shown in Appendix C. The EAA can
measure particles in the size range 0.0075 to 0.6 ym; the Royco measures
particles in the size range of 0.3 to 10 ym.
*Note: Throughout this paper "EAA" is an acronym for the Model 3000 Whitby
Electrical. Aerosol Analyzer and not for the Model 3030 Electrical Aerosol
Analyzer cis it usually is. Although both the Model 3000 and 3030 instru-
ments operate on the same basic principle, there are important differences
in design and presumably, also in performance (personal communication,
B. Y. H. Liu to G. J. Sem).
20
-------
Photograph 5,
Instruments, at the left an Environment I condensation
nuclei counter (CNC) on top of a voltage recorder used
to record the Model 3000 Whitby Electrical Aerosol
Analyzer (EAA) in the center and at the right a Model
200 Royco Optical Particle Counter.
Aerosol generation times of 5 to 30 minutes were used and the particle
counts measured during this time were averaged for each size range. Care was
taken to achieve the proper dilution of the generated aerosol so that the
particle concentrations of the smaller particles were not high enough to
cause instrument saturation and also that the particle concentrations of the
larger particles were significantly above the background counts. Laboratory
air was used to dilute the generated aerosols. On most days the background
(laboratory air without generated aerosol) particle counts were less than 10%
of the diluted generated aerosol particle counts. Because the laboratory is
located in the desert, the air is almost always very dry and there was no
problem of water condensation even when only one diluter stage was used.
The instruments measured diluted aerosol particles that were generated 9
seconds prior to entering these instruments.
: ;
-------
SIZE DISTRIBUTIONS
Most size distributions are plotted as, N(D) , the total number of particles
per cm3 with diameters greater than D vs. particle diameter, D. The outputs
from the ROYCO optical particle counter are plotted in this format without
modification. The outputs from the electrical aerosol analyzer (EAA) are
added to the ROYCO count at D=0. 6 ym to calculate N(D) for D's smaller than
0.6 ynu At particle sizes between 0.3 and 0.6 ym the outputs from both
instruments are shown. Some of the size distributions are plotted as AN/AlogD.
Aerosol particle volumes were calculated assuming that the particle
counts s AN^, in each size interval (D-^ to D-j^) are the numbers of single
spherical particles of diameter equal to the geometric diameter of the size
interval (square root of DiD-j^). Thus the volume of particles, AV, counted
in size Interval V, D- is
The volurie distributions are plotted as volume %. The EAA counts are used
for sizes smaller than 0.6 ym; Royco counts are used for the larger sizes.
22
-------
SECTION 6
RESULTS
Measurements, Observations and Comparisons
THE PARTICLES
The aerosol particles were measured approximately 9 seconds after generation.
They remain suspended in air while they are sized and counted by the EAA and
Royco instruments. They do not, however, remain as single spherical particles,
even though most of them probably formed as single, "nearly" spherical particles.
Most of the single particles quickly coagulate with other particles. Fuchs
(1964, pp. 288-315) offers a good introduction to the theory of coagulation
of aerosols. Readers not familiar with the subject can better understand our
experimental results by studying this or similar works.
Several of the generated aerosols were precipitated and photographed.
The larger particles (D>10 ym) were sampled by letting them impact a glass
slide coated with adhesive; the smaller particles were thermally precipitated
onto SEM and TEM grids.
Photograph 6a shows a 25 ym diameter zinc oxide particle covered with
smaller particles which coagulated with it. Photograph 6b shows a 20 ym
diameter hollow iron oxide (probably FesOi,) particle. Photograph 6c shows a
close up of the surface of an iron oxide particle with some smaller captured
particles. Photograph 6d shows another hollow particle captured while generat-
ing the electric arc furnace simulation aerosol. Photograph 7a shows a large
number of iron oxide particles several of which have holes showing that they
are hollow; possibly all of these particles are hollow. Further evidence of
large hollow particles was noted by observing the behavior of particles which
precipitated onto the stack. After a test when the stack cooled, there was
often a slight draft; particles dislodged by rapping the stack floated up and
out of the stack. After one test where MgO aerosol had been generated,
particles, later identified as clusters of three to six 50 ym diameter
particles, were observed floating up. The air velocity was less than 0.2
m/s. This indicates a maximum particle density of 2.5 g/cm3 which is only
68% of the density of MgO.
When calcium nitrate solutions were burned, the captured particles
appeared to have "wet" smooth surfaces. We were not able to identify the
chemical composition with X-ray diffraction analysis. Photograph 7b shows
one of these hydrated particles.
23
-------
Photograph 6a. Zinc oxide particle Photograph 6b. Iron oxide particles
coated with submicron particles. from basic oxygen furnace simulation
aerosol BB.
Photograph 6c.
surface.
Iron oxide particle
Photograph 6d. Hollow particle from
the electric arc furnace simulation
aerosol.
-------
1 m f* •
lP^/5*
#> „ *v .'
-" ^^ *
1-
Photograph 7a. A large number of
iron oxide particle several of which
have holes showing that they are
hollow.
Photograph 7b. A hydrated particle
containing an unidentified calcium
compound.
The smaller particles are seldom seen as single spheres. Photograph 8a
shows a single 0.2 ym particle with several particles of about 0.01 y diameter
coagulated with it. Photograph 8b shows a wide spectrum of particle sizes
and aggregates. The particles on these plates were captured from the coal
fly ash simulation aerosol. Photographs 8c an 8d show particles captured
from the electric arc furnace simulation aerosol; they are primarily chain
aggregates of 0.01 to 0.02 ym particles. For most of the aerosols generated,
the bulk of the aerosol mass appears to exist as chain aggregates. The
aggregates form from the very high concentrations (>10 particles per cm )
of 0.01 to 0.02 ym particles. The Royco and EAA are calibrated with spheri-
cal particles with specific physical properties. Consequently, the indicated
particle size ranges may be biased to various degrees by different aerosols.
The Royco and EAA counts are, in reality, a measurement corresponding to the
number of spheres of a standard material that could produce an equivalent
signal. The aggregated particles have non-ideal shapes and it is difficult,
if not impossible, to calculate their electrical, aerodynamic and optical
properties.
PARTICLE COUNTS
The size distributions of 26 different aerosols were measured simultaneously
by the EAA and Royco instruments. In many of these tests the CNC instrument
was used also. There are some interesting disagreements between the instrument
counts. The particle counts were used to compute aerosol volume concentrations
and these were compared with concentrations determined from the solutions and
burn times. Again there were some interesting disagreements.
-------
l.i f'
r
0.1
0.5 u
Photograph 8a. Few of the submicron Photograph 8b. Particles from the
particles were seen as single spheres. coal fly ash simulation aerosol.
0.1
Photograph 8c. Particles from the
electric arc furnace simulation
aerosol.
Photograph 8d. Particles from the
electric arc furnace simulation
aerosol with more magnification.
26
-------
Figure 6 shows the background particle counts for a typical test with
three stage dilution. These counts were recorded prior to burning the solution.
To measure a test aerosol, the particle counts in each size range must be
high enough so that the background counts can be subtracted to yield a signi-
ficant difference. On most days the background counts were steady; when they
were not, a larger difference between background and burn was required.
Since the variability of particle counts increases as the number of particles
counted decreases, the ratio of burn counts to background counts must be
greater for the larger sized particles. The most likely variable which would
reduce the reproducibility of the following aerosol size distribution measure-
ments was a change in the background particle concentration during a test.
By carefully controlling this, the reproducibility of the measurements was
very good.
Q
Z
1010
108
107
10"
10*
EAA
COUNTS
DIAMETER,
Figure 6. Typical background measure-
ments before aerosol generation. The
background measurements shown in this
figure are multiplied by 6900 because
with three stage dilution the equiva-
lent of 6900 parts of air is mixed with
one part of aerosol.
27
-------
Th<> generated aerosol characteristics depend on the rate at which solution
is burntid in the generator. The flow rate of solution sprayed into the gener-
ator is dependent on the solution viscosity. The viscosity is dependent on
the solution temperature at the nozzle; thus, the generated aerosol character-
istics depend on the solution temperature at the nozzle. Both the temperature
of the solution in the pressure tank and of the air entering the generator
affect the solution temperature at the nozzle. The magnitude of this effect
was not measured. We believed, however, that this effect was not important
for the 26 aerosols generated, measured and described in this report.
For most tests three stage dilution was required to dilute the aerosol
particle concentration to prevent Royco and EAA saturation. Three stage
dilution reduces the particle concentration by a factor of 6,900. Thus, with
three stage dilution, an undiluted aerosol particle concentration of 200
particles per cm3, is reduced such that less than two particles are counted
by the Eoyco in a 0.3 minute counting time. The concentration of aerosol
particles with diameters greater than Dx is roughly inversely proportional to
the cube of Dx. With three stage dilution the larger particles had low
concentrations. An excessively long time (several minutes) would have been
required to count a representative number of them to determine their true
concentration. This causes an effective size cut-off. The concentrations of
large particles are best measured with two stage dilution.
For some aerosols the particle counts decreased in the size range 0.015
to 0.03 ym and then increased again in the size, range 0.0075 to 0.015 ym.
This may have been caused by our using the Whitby-Cantrell data reduction
constants.
Because of coincidence counting of particles by the Royco, all Royco
counts of diluted aerosol with concentrations in excess of 100 particles/cm3
were assumed in error (saturated) and are not reported. In some cases, even
with three stage dilution, one or both of the instruments was saturated for
the smallar sized particles (e.g., Royco 0.3 to 0.5 ym), and additional
dilution of the aerosols had to be used. This was accomplished by using
smaller orifices at the diluter fume outlet of stages 1 and 2 (see G in
Figure 5). This increased dilution is defined as "non-standard three stage
dilution". MgO and SiOa aerosols were used to determine the increase in
dilution, The amount of dilution was found to be a function of particle
size; the dilution slowly increased with increasing particle diameter.
Average values of dilution were used to convert the instrument particle
counts to undiluted aerosol particle concentrations. With non-standard three
stage dilution, the average dilution in the EAA size range (0.0075 to 0.3 ym)
is 38,700 and in the Royco size range (0.3 to 4 ym) 65,700.
App€;hdix B has figures showing the size distributions of most of the 26
different: aerosols which were generated and measured but not shown in the
text. The figures are plotted as N(D) vs D. Appendix C has the particle
concentrations and volume percents in each size interval for the 26 different
aerosols.
28
-------
Mass Balances
The solution composition, flow rate, aerosol stack exit velocity and
temperature were known for each experiment, therefore the aerosol particle
loadings (g/m3) could be determined by assuming that all of the solid products
of combustion were suspended as aerosol particles and an insignificant portion
collected onto the flame cone, flame holder and stack. With one exception
(the 1.8 g/m3 basic oxygen furnace simulation aerosol), this assumption is
good for all of the aerosols reported. In order to calculate the volume
concentrations from the particle loadings, it was necessary to assume particle
densities. Handbook values of densities of those compounds identified by X-
ray diffraction analysis were used.
The volume concentrations were also determined by using the particle
counts from the instruments. To do this, two assumptions were made: first,
that the particles counted in the interval Dx»Dy all have a diameter DQ
equal to the square root of DxDy (geometric mean diameter); second that the
particles are all solid spheres. For most of the aerosols these assumptions
are in error, especially the second as photographs 6a-8d show.
The Royco and EAA instruments are calibrated with a dilute aerosol of
spherical particles. The particle counts from these instruments were used to
determine particle volumes and solids loadings of aggregated aerosols. It is
not surprising that the volumes and loadings sometimes disagreed by an order
of magnitude with the volumes and loadings determined from solution composi-
tions and flow rates (generation rates) .
For many of the aerosols, the bulk of the measured volume was contained
in particles with diameters less than 0.6 ym (sizes measured by the EAA).
For many of these aerosols, the volumes determined from the EAA particle
counts were larger than the volumes determined from the generation rates.
Tests were conducted with SiOa, MgO, ZnO and FeaOs aerosols to see if
the loadings determined from the generation rates and those from the particle
counts would be in better agreement at lower particle loadings. Table 8
shows aerosols with two loadings of SiOa, one with that of the coal fly ash
simulation aerosol (0.163 cm3/m3), another with a much lower loading (0.004
cm3/m3). Table 9 shows MgO, ZnO and FeaOs at higher and lower particle
loadings. Figure 7 shows the particle size distribution of these two SiOa
aerosols as measured by the EAA. The following points seem to show that the
EAA was not saturated at either SiOa loading. The maximum current on the EAA
was 6.9xlO~12 amperes on stage 1 for the SiOa aerosol containing the higher
loading; the EAA did not normally saturate at this current. The ratio of the
volume determined from the EAA particle counts to that determined from the
generation rates was not reduced at the lower Si02 loading; in fact, it in-
creased a little (see Table 8) . These observations indicate that instrument
saturation is not responsible for the loadings determined from the EAA particle
counts being higher than those determined from the generation rates. Lowering
particle loadings did not lead to better agreement.
Since the only solid product of combustion of (CHs^SitOCaHs^ is Si02,
there is nothing in the chemistry to indicate that the loadings determined
29
-------
TABLE 8. AEROSOL PARTICLE LOADING (Weight or Volume/Hot Gas Volume)
Measured in 0.425 m Stack with Three Stage Dilution
00
o
Aerosol
Particle
Composition
Si02
Si02
Si02 & A1203-2H20
Si02, A1203'2H20 &
Fe203 & Fe30i,*
Same composition
as above*
Same as above
with Li2C03
Fe203 & Fe30i,
Fe203 & Fe30i»
A1203'2H20, Fe203 & Fe30i»
A1203»2H20
Determined From
Generation Rate
g/m3 cm3/m3
0.36
0.009
0.62
0.66
0.66
0.94
0.15
0.12
0.41
0.26
0.16
0.0039
0.26
0.25
0.25
0.39
0.029 ,
.024
0.13
0.1
rarcicjue Apparent
Determined Particle Density Maximum
From EAA & Loading (Handbook Particle
Royco Counts (Counted/ Value) Density
cm3/m3 generated) g/cm3 g/cm3
1.0
0.027
1.4
1.1
1.1
.55
0.027
.009
0.08
0.02
6.0
7.0
5.0
4.0
4.0
1.4
0.9
0.4
0.6
0.2
2.2
2.2
2.34
2.6
2.6
2.43
5.17
5.17
3.06
2.55
0.37
0.31
0.47
0.65
0.65
' 1.7
-
-
-
-
*These aerosols are the coal fly ash simulation aerosol.
-------
TABLE 9. AEROSOL PARTICLE LOADING (Weight or Volume/Hot Gas Volume)
Measured in 0.25 m Stack with Three Stage Dilution.
u>
Aerosol Particle
Composition
Si02
MgO
MgO
MgO
ZnO
ZnO
Fe203**
Fe203**
Fe203**
Fe203 & FeaOi,
MnaOi,
Cr203
Ca(OH)2-xH20
Ele Arc Sim. Aero.
Ele Arc Sim. Aero.
Above w/o Mn
Determined From
Generation Rate
g/m3 cm3/m3
0.096
0.62
0.054
0.49
0.087
0.94
0.071
0.56
6.0
0.78
0.62
0.69
0.29
0.65
1.2
1.0
0.043
0.17
0.015
0.13
0.016
0.17
0.014
0.11
1.17
0.15
0.13
0.13
0.13
0.14
0.24
0.21
Determined
From EAA &
ROYCO Counts
cm3/m3
0.544
3.6
0.35
2.9*
0.18
1.2*
0.17
0.89*
22.*
0.054
0.11
0.95
1.1
0.054
0.064
0.072
Particle
Loading
(Counted/
Generated)
13
21
24
22
11
7
12
8
19
0.36
0.85
7
8
0.40
0.26
0.34
Particle
Density
(Handbook
Values)
g/m3
2.2
3.65
3.65
3.65
5.47
5.47
5.12
5.12
5.12
5.14
4.72
5.21
2.24
4.79
4.79
4.81
Apparent
Maximum
Particle
Density
g/cm3
0.17
0.17
0.15
0.17
0.49
0.78
0.43
0.64
0.27
-
-
0.74
Hydrated
-
-
-
* Used non-standard 3-stage dilution (see text)
** From Fe(CO)5
-------
io">
109
108
107
106
105
103
102
101
0.01
-0.163 cm3/m3
.0.004 cm3/m3
0.10
10.00
DIAMETER, j
Figure 7. Comparison of measurements
with the electrical aerosol analyzer
(EAA) of two SiOa aerosols, one con-
taining 0.163 cm3 of SiOa per m3 of
aerosol, another containing 0.004 cm3/
m3. The EAA does not appear to have
been saturated.
from the; generation rate are in error. When alcohols (methyl, ethyl or
propyl) or propane were burned in the generation apparatus with three stage
dilution;, there was a negligible increase in particle counts over the back-
ground counts.
Tables 8 and 9 show several aerosol particle loadings as determined
from the; generation rates and from the instrument particle counts. The
aerosols were generated in the 0.42-m stack (Table 8) and in the 0.25-m
stack (Table 9). The "apparent maximum particle density" equates the particle
loadings determined from the instrument particle counts (using the assumption
that the particles are solid spheres) to those from the generation rates.
The apparent maximum particle densities shown in Tables 8 and 9 are equal to
the handbook values of densities divided by the counted/generated particle
loadings.
32
-------
Many of the aerosols in Table 9 have nearly constant volume loadings.
This was purposely done so that an aerosol of one chemical compound could be
compared with an aerosol of another compound. The volume loadings were
based on handbook values of densities. The small variation in particle
loadings was caused by small variations in the solution flow rates which not
only changed the solids generation rates but also the aerosol gas flow rate.
This variation in solution flow rates was caused by varying solution viscosities
due primarily to the different chemical compositions of the solutions. The
volume loadings varied from 0.13 to 0.17 cm3/m3. The MgO, ZnO and Fe203
aerosols were also generated with volume loadings of only 10% of the constant
volume (see Table 9). The Si02, MgO, ZnO and Fe203 (from Fe(CO)5) aerosols
were measured in much the same manner with the EAA.
CNC vs. EAA Counts
Table 10 shows the total particles counted by the Environment I CNC*
and by the EAA and Royco. Since the number of particles counted by the
Royco was small compared to the number counted by the EAA, Table 10 essen-
tially shows a comparison of the EAA and CNC counts. For most aerosols the
particles counted by the EAA were between 5 and 6 times as great as those
counted by the CNC. The average value of the ratio of concentrations (EAA/CNC)
is 6.1 with a variance of 3.5. The Fe203 from Fe(CO)5 and H2SOil'H20 aerosols
were not included in the computation of an average value because they appear
to behave differently from the other aerosols. The best agreement between
the CNC and EAA was for the aerosol containing H2SOtj*xH20 droplets; perhaps
with the other aerosols, particle response as condensation nuclei was slower
and growth did not occur on all of the aerosol particles.
EAA vs. Royco Counts
For some aerosols the particle counts from the EAA and Royco in the
size range where they overlap (0.3
-------
TABLE 10. AEROSOL PARTICLE CONCENTRATION (CNC vs EAA)
Measured in 0.25 m Stack Using Three Stage Dilution
OJ
Aerosol
Particle
Composition
Si02
Si02*
MgO
MgO
MgO**
ZnO
ZnO**
Fe203
Fe203**
Fe203**
Fe203 & Fe30i,
Mn30i»
Cr203
Ca(OH)2
Ele Arc Sim Aero.
Ele Arc Sim Aero.
Above w/o Mn
H2SO.»»xH20
Coal Fly Ash Sim Aero.*
Above with Li*
Aerosol
Generation
Rata
g/m3
0.096
0.36
0.62
0.054
0.49
0.087
0.94
0.071
0.56
6.0
0.78
0.62
0.69
0.29
0.65
1.2
1.0
Very low
0.66
0.94
Particle
From
CNC
(///cm3)
3.4E8
9.3E7
1.7E8
1.3E8
1.2E8
2.2E8
1.4-1.7E8
2.4E8
0.2-1.0E9
4-30E8
2.6E8
2.4E8
2.1E8
3.2E8
3.0E8
2.4E8
2.8E8
1.0-1.3E9
1.1E8
1.1E8
Cone en t r a t ion
From
EAA
(#/cm3)t
1.9E9
9.4E8
EAA saturated
7.3E8
1.1E9
1.7E9
7.8E8
1.E9
2.3E9
2.2E10
7.8E8
1.3E9
1.0E9
1.5E9
1.5E9
1.3E9
1.3E9
2.4E9
9.2E8
7.4E8
Rati'.O of
Concentrations
(EAA/CNC)
5.6
10
-
5.6
9.2
7.4
5.8
6.3
2.3-12
7.3-55
3.0
5.6
4.8
4.8
5.0
5.5
5.4
2.1
8.4
6.7
t The EAA counts of all particles larger than 0.0075 ym
* Generated in 0.42 m Stack
** Non-standard Three Stage Dilution
-------
TABLE 11. AEROSOL PARTICLE COUNTS. Measured in 0.42 m
Stack with Three Stage Dilution
Aerosol Particle
Composition
Si02
Si02*
Si02, A1203-2H20
Si02, Fe203 & Fe30^
Si02, A1203«2H20 &
Fe203 & Fe30it
Same as above
Same as above with Li2C03
Fe203 & FesOn
Fe203 & Fe30it
A1203'2H20, Fe203 & Fe304
A1203-2H20
Approximate
Aerosol
Generation
Rate g/m3
0.36
0.009
0.62
0.51
0.66
0.66
0.94
0.15
0.12
0.41
0.26
Diameter of
Largest
Particle
Counted
(Vim)
.6
0.6**
4
3
3
8
3
1.5
1.5
8
3
Particle
Volume
Median
Diameter
(ym)
0.2
0.06
0.2
0.2
0.2
0.2
0.2
0. 05 at
0.05
1.2
>1.5
EAA
Counts
.3
-------
CO
TABLE 12. AEROSOL PARTICLE COUNTS
Measured in 0.25 m Stack with
Three Stage Dilution
Aerosol Particle
Composition
Si02
MgO
MgO
MgO
ZnO
ZnO
Fe2o3**
Fe203**
Fe203**
Fe203 & FesOi,
Mn30i,
Cr203
Ca(OH)2»xH20
Ele Arc Sim. Aero.
Ele Arc Sim. Aero.
Above w/o Mn
H2SO«t'xH20
Aerosol
Generation
Rate g/m3
0.096
0.62
0.054
0.49
0.087
0.94
0.071
0.56
6.0
0.78
0.62
0.69
0.29
0.65
1.2
1.0
Very Low
Diameter of
l.arooot-
0
Particle
Counted
(ym)
0.6
6
4
4*
3
4*
0.6
0.8*
8*
3
3
10
6
3
4
4
0.06
Particle
TT 1 — —
V w J.U1I1C
Median
Diameter
(ym)
0.15
-
0.2
0.3
0.1
0.2
0.1
0.2
1.05
0.3
0.4
>4.7
>3.6
0.08
0.2
0.3
<0.015
EAA
Counts
.3
-------
10'° -
10s -
108 -
10' -
I "H
z
103 -
104 -
103
0.01
EAA COUNTS
10'
10'°
109
10"
a
| 106
z
10=
0.10
1.00
Figure 8. Comparison of measurements
with the Royco optical particle
counter and the EAA of the 0.78 g/m3
FezOs & FesOi4 aerosol. The measure-
ments with the Royco & EAA are in
close agreement.
10"
102
0.01
-EAA COUNTS
0.10
1.00
10.00
DIAMETER, >im
Figure 9. Comparison of measurements
with the Royco and EAA of the 0.054
g/m3 MgO aerosol. The measurements
with the Royco and EAA do not agree
with each other.
red iron oxide aerosols
the ratio was as high as 30.
EAA to Royco counts.
generated from iron pentacarbonyl (Fe(CO)5),
Si02 and MgO aerosols had very high ratios of
The EAA and Royco instruments also did not count the same number of
particles in the size range 0.3
-------
VOLUME DISTRIBUTIONS
Tabi.es 11 and 12 show the diameters of the largest particles counted and
the voluiie median diameters. The Si02 aerosols had very few if any particles
larger than 0.6 ym. Fe203 aerosols generated from Fe(CO)5 also had few, if
any, paruicles larger than 0.8 ym except at the very high solids loading (6
g/m ). Both of these aerosols (Si02 and FeaOa) were generated from liquids
with high vapor pressures. All of the nitrates decomposed to form aerosols
with high enough concentrations of particles larger than a 1 ym to be counted
by the Royco with three stage dilution. Even dilute solutions containing
nitrates yielded some super ym particles (see the MgO, 0.054 g/m3 and ZnO,
0.087 g/n.3 aerosols in Table 12).
For most of the aerosols generated, the volume of those particles with
diameters between the cut-off (concentrations too low to count) and 10 ym was
probably a small amount relative to the measured volume because the volumes
of the particles with diameters larger than 1 micron decreased rapidly (see
Figure 10). However, the A1203»2H20, Ca(OH)2*xHaO and Cr203 aerosols have a
large portion of their measured volume contained in large particles (see
Figure 11). Volumes contained in particles larger than 10 ym were not measured
because the Royco Model 200 does not size particles larger than 10 ym. Also
the sample train (diluter and associated plumbing) was not designed for
particles larger than about 10 ym.
When 1.8 g/m3 of FeaOi* with 5% Ca(OH)2*xH20 aerosol was generated, only
10% of the generated volume was accounted for by the instrument particle
counts. The volume distribution is shown in Figure 12a. For this aerosol
and for some of the others, there was probably a mode in the volume distri-
bution conprised of particles too large to measure with the instruments. This
indicates (that a limited amount of nitrate can be added to a given volume of
ethanol; adding more than this limit results in a significant portion of the
generated aerosol particles being too large to qualify as fine particulates.
The 1.8 g/m3 Fe30it aerosol was probably an example of this.
The 1.8 g/m3 FesO^ aerosol was sampled by letting it impact a glass
slide coated with adhesive. This method is probably satisfactory for obtain-
ing a representative sample of particles larger than 10 ym if the density of
the hollow particles is not too low. Photographs 6b and 7a show SEM photo-
graphs of samples of this Fe30it aerosol. One photograph was used to measure
and count 494 particles. The volume distribution from this work is shown in
Figure 12b . As is often the case in particle size distributions with less
than a few thousand particles counted, not enough large particles were counted.
Twenty-nine percent of the volume was contained in two particles, one 42.5 ym
and the other 70 ym.
Some of the aerosols had a bimodal volume distribution in the measured
size range (0.0075 to 10 ym) . Figure 13 shows the bimodal size distribution
of the A1203»2H20, Fe203 and Fe30i» aerosol (0.41 g/m3).
38
-------
60
54
48
42
36
30
24
18
12
0.01
0.10
DIAMETER,
1.00
10.00
Figure 10. The volume distribution determined from EAA and Royco
measurements of the 0.62 g/m3 MnaOi, aerosol. Most of the volume
of aerosol particles appears to be included in the size range
measured by the instruments.
o
60
54
48
42
36
30
24
18
12
0.01
0.10
DIAMETER, ion
1.00
10.00
Figure 11. The volume distribution determined from EAA and Royco
measurements of the 0.69 g/m3 Cr203 aerosol. A significant portion
of the volume of aerosol particles probably exists as particles
larger than 10 ym.
39
-------
54
48
42
36
30
24
18
12
6
0
1
-
0.01
0.
I
10
DIAMETER,
-
1.00
/urn
-
-
-
-
-
10.00
Figure 12a. The volume distribution determined from EAA & Royco
measurements of the 1.8 g/m3 FesOit aerosol. This aerosol was
found to have a volume mode of larger (D>10 ym) particles.
60
54
48
42
36
30
24 -
18 -
12
0.1
10
100
DIAMETER, tan
Figure 12b. FesOi, aerosol 1.8 g/m3, the volume distribution of some
large particles captured on a glass slide. Possibly 80-90% of the
aerosol mass existed as large hollow particles.
40
-------
60
54
48
42
36
30
24
18
12
0.01
0.10
DIAMETER, ion
1.00
10.00
Figure 13. The volume distribution determined from EAA & Royco
measurements of the 0.41 g/m3 A1203 2H20, Fe203 and FeaOn aerosol.
This aerosol has a bimodal distribution.
PARTICLE COMPOSITIONS
Table 13 shows the X-ray diffraction results. Compounds (usually
nitrates) were dissolved in alcohol, sprayed and burned. The hot products
of combustion rose vertically in the 25.4 cm diameter stack. The solid
particles suspended in the fume were collected at the top of the stack by
placing the rectangular nickel plated brass plate at a 45 angle at the top
of the stack. The amount of collected material varied from a few milligrams
to about a gram. Burn times for collection were about 6 minutes. A part of
each sample was immediately sealed in a glass capillary tube to await X-ray
diffraction analysis. X-ray diffraction analysis did not yield the composi-
tion of those aerosols with amorphous particles.
41
-------
TABLE 13. PARTICLE COMPOSITION
Compound
(grams)
Fe(N03)3'9H20*
586g
Fe(N03)3-9H20
757g
Ca(N03)2*4H20
33g
Solvent (s)
(liters)
Ethanol
1.55,
Ethanol
21
Particle
Composition
(X-ray diff.)
Fe203 with a trace
of Fe30i»
Fe30it with some
Fe203
(calcium not detected)
Mn(N03)2'6H20
lllg
Cr(N03)3-9H20
160.5g
Mg(N03)2-5H20
400g
Mg(N03)2-5H20
200g
Zn(N03)2-5H20
468g
Zn(N03)2'6H20
362g
Cd(N03)2-^H20
lOlg
VOSOit'xH2()
69g
Cd(N03)2'^H20
50.5g
VOSCVxH20
69g
Ethanol
1.5A
Ethanol
1.5SL
Ethanol
2H
Ethanol
21
Methanol
2H
40cc water
Ethanol
21
Methanol
2£
200cc water
Methanol
U
lOOcc water
Methanol
U
lOOcc water
(hausmannite)
Cr203
Face centered cubic
MgO (periclase)
Same as above
Hexagonal ZnO
(zinkite) plus an
unknown organic
compound
Hexagonal ZnO
CdO plus
Cd(OH)N03H20
V02,V201|,
and V205
Primarily
some CdO,
2CdO«V205, and
V02,V20!», and
V205
*The Fe(NOs)3-9H20 was exposed to the ambient air for two
mixing into solution. Both the Fe(N03)3"crystals and the
reddish brown and probably already contained Fe203.
days prior to
solution were
42
-------
COAL FLY ASH FUME SIMULATION AEROSOL
Table 14 shows several characteristics of the coal fly ash fume simula-
tion aerosol.
TABLE 14. CHARACTERISTICS OF THE COAL FLY ASH
SIMULATION AEROSOL
Stack: Diameter, 0.425m; height, 3.04m
Fume Exit Temperature: 450°K (177°C)
Fume Exit Velocity: 5 m/s
Fume Flow Rate: 0.7 m3/s, 0.51 kg/s
Particle Loading: 0.66 g/m3 @ 450°K; 1.1 g/m3 @ STP
Particle Mass Median Diameter: 1 ym from Celesco impactor;
Volume Median Diameter: 0.2 ym from EAA & Royco
Condensation Nuclei: l.lxlO8 particles/cm3
Nominal Chemical Composition (wt%): 50% Si02, 30% A1203 and
20% Fe203
Color of Particles: most are pink, some are black
Bulk Density: about 0.3 g/cm3; packed density: 0.9 g/cm3
Electrical Resistivity: 2xlO!1 - 2xl012 ohm cm @ ambient temp.
Chemical Analysis
SEM with X-ray Non-Dispersive Analysis
Individual aerosol particles were analyzed using a SEM with an ETEC
(Princeton Gamma-Tech) X-ray non-dispersive analysis system to determine if
elements Si, Fe and Al existed separate from each other or were bound together
in the same particle. Since the beam penetrates about 0.5 ym into the sample
or substrate, only particles larger than 0.5 ym diameter could be sampled
without having excessive background signal. Ten particles were sampled. One
particle contained Si, three contained all three elements, and six of the
particles contained Si and Al, but not Fe. This is encouraging since coal
fly ash usually contains Si and Al found together as mullite (3Al2Os«2Si02).
Most of the particles had more Al present than Si; however, the Al to Si
ratios were different for each particle. Analysis of an area on the graphite
pedestal without any particles visible on the SEM indicated that perhaps
there was a fine "coating" of small particles containing Si. These were too
small to be seen with the SEM on the graphite substrate. The background
signal was high for this measurement. Analysis of a sample which collected
on the flame cone and did not flow up the stack showed that the sample con-
tained Al as the major component; next in abundance was Fe; the Si content
was about 10% that of Al.
43
-------
Cyclotron Excitation
The coal fly ash simulation aerosol was sampled for one minute with the
Lundgren Impactor (Model 4220 operating at a flow rate of 4.72X10"1* m3sec
(4 CFM)) using the first stage of the diluter to cool and dilute the fume.
An 0.8 ym Nuclepore filter was located down stream from the Lundgren Impactor.
The solies loading on the filter was so high, the flow rate dropped 25%
during the sampling time. The collected samples were analyzed by Crocker
Nuclear laboratory (University of Calif., Davis) using cyclotron excitation.
Table 15 shows the results. The samples were also examined with an optical
microscope. The particle sizes on the stages overlapped each other (see
Table 15). Stages 1 and 2 had more uniform particles than stages 3 and 4.
Many of the particles on stages 3 and 4 were nonspherical agglomerates of
smaller particles. On stages 1-3, there was good separation of the captured
particles indicating little likelihood of particle bounce-off. The particles
on the Nucleopore filter were stacked on top of one another and it was not
possible to determine sizes.
TABLE 15. Cyclotron Excitation Analysis of Lundgren Impactor
and Nuclepore Filter Samples of the Coal Fly Ash
Simulation Aerosol.
Stage
1
2
3
4
Filter
Total Wt
Wt.%
Wt.% as
Vol% as
Theoretical
Dia. Range
Spheres of
Density
2 g/cm3
(Um)
1.2-140
3.5-12
1.4-3.5
00 35-1. 4
<0.35
oside
oxide*
Vol% stages 1-3*
Vol% on
filter*
Observed Wt.
Diameter Al
Range ng/
(ym) cm2
>7 730
2-7 4,500
1-3 3,000
<1-10 2,600
40,000
51,000
27
30
35**
75**
31**
Wt.
Si
ng/
cm2
140
120
860
2,500
103,000
107,000
57
59
60
11
65
Wt. Total
Fe Wt.
ng/ ng/
cm2 cm2
100 970
3,600 8,200
2,500 6,400
2,600 7,700
21,000 164,000
30,000 188,000
16
11
5
14
4
Wt.%
0.5
4.0
3.0
4.0
88.0
100
* Handbook values of the densities were used to calculate particle volumes.
**As A1203'2H20
44
-------
Eighty-eight percent of the sample mass was located on the filter indicat-
ing the small equivalent aerodynamic diameter of most of the simulation
aerosol particles. Some of the Fe2C>3 appears to be lost in the stack or
diluter. The nominal wt% generated, based on the weights of nitrates in the
basic solution, is 50% Si02, 30% A1203 and 20% Fe203; the collected samples
had 59% as Si02, 30% as A1203 and 11% as Fe203. The wt% of oxides on the
filter sample was 68% as Si02, 23% as A1203 and 9% as Fe203.
X-ray Diffraction
X-ray diffraction analysis showed the bulk of the particles to be amorphous.
For example, after exposure to X-rays for 48 hours, one sample collected on
the flat plate did not yield a pattern. Another sample collected on a 0.6 cm
diameter stainless steel rod with its longitudinal axis normal to the flow
had a faint pattern believed to be from a mixture of iron oxides. This
sample was brown however and the bulk of the aerosol particles are pinkish.
When the oxides of each element (Si, Al and Fe) were generated separate from
the others, samples of the oxides of Si and Al were amorphous. The iron
oxide sample consisted of crystalline Fe30i, and Fe203.
Because the particles are amorphous, the X-ray diffraction analysis of
the coal fly ash simulation aerosol did not conclusively show what compounds
the Si, Al and Fe form.
Size Distribution
Figures 14 (plotted as AN/Alog D vs. D) and 15 (plotted as N(D) vs. D)
show the size distribution of the coal fly ash simulation aerosol generated
on 22 November 1976. Figure 1.6 shows the size distribution of the coal fly
ash simulation aerosol generated again on 3 December 1976. Figures 15 and 16
show the reproducibility when solutions with the same chemical composition
are burned under the same conditions on different days. Figures 17 and 18
show the volume distribution of the coal fly ash simulation aerosol (generated
on 22 November and on 3 December); again the reproducibility is good.
The particle counts and size distribution were reproducible from run to
run, but the volume of particles counted did not agree with the mass of
particles generated. The apparent mass determined from the particle counts
was four times the generated mass, suggesting the presence of hollow particles,
aggregates with densities differing from the handbook values or an instrumen-
tal error. Because the EAA and Royco did not count the same number of
particles in the diameter range 0.3 to 0.6 ym where their particle counts
overlap, there definitely was an instrumental "error" in that size range.
Dr. Raymond L. Chuan (California Measurements Inc., Sierra Madre, CA)
came to China Lake and measured the mass distribution. He used a ten stage
Celesco C1000 cascade impactor equipped with quartz crystal sensors (Chuan,
1976). Table 16 shows a comparison of his measurements and those of the EAA
and Royco. The measured solids loading using the impactor accounted for 42%
of the generated solids, which is more reasonable than the 400% obtained with
the ROYCO-EAA measurement system. A particle density of 2 g/cm3 was used to
45
-------
10"
1010
109
108
107
106
105
104
103
102 -
-EAA COUNTS
I
0.01
0.10 1.00
DIAMETER, //m
10.00
Figure 14. The size-distribution,
plotted as AN/AlogD vs. D, from measure-
ments with the EM and Royco of the coal
fly ash simulation aerosol generated
22 November 1976.
determine the particle sizes collected on each impactor stage and to deter-
mine the mass of the particles counted by the ROYCO and EAA. The handbook
value of density is 2.6 g/cm3.
An examination of Table 16 shows that the measurements of the Celesco
cascade impactor are in fair agreement with the ROYCO optical particle counter
but in very poor agreement with the EAA.
Inasmuch as 88% of the aerosol passed through the Lundgren impactor
without impacting any of the stages (see Table 15), the EAA and Lundgren
measurements agreed with each other.
Table 17 gives a comparison of the measurements of the Celesco and
Lundgren irnpactors. The particle mass median diameter determined from the
Lundgren data is less than 0.35 ym; from the Celesco data 1.05 ym (see
Figures 19 and 20). The volume median diameter determined from the ROYCO-EAA
data is 0.23 ym (see Figure 20).
-------
101"
nio
109
109
106
106 -
10'
|5 -
104 -
10* -
102 -
102 -
DIAMETER, jim
0.10 1.00
DIAMETER, /jm
Figure 15. The size distribution from
measurements with the EAA and Royco of
the fly ash simulation aerosol gener-
ated 22 November 1976.
Figure 16. The size distribution
from measurements with the EAA and
Royco of the fly ash simulation
aerosol generated 3 December 1976.
-------
so
54
18
s«
LiJ
3
i6
30
24
18
0.01
0.10
DIAMETER,
1.00
10.00
Figure 17. The volume distribution determined from the EAA and
Royco measurements of the fly ash simulation aerosol generated
22 November 1976.
liO
154
'.8
42
56
50 -
18 -
12 -
0.01
0.10
DIAMETER, /am
1.00
10.00
Figure 18. The volume distribution determined from the EAA and
Royco measurements of the fly ash simulation aerosol generated
3 December 1976.
-------
TABLE 16. MEASURED MASS DISTRIBUTION OF THE
COAL FLY ASH SIMULATION AEROSOL
(0.66 g/m3 of aerosol is generated)
Celesco Cascade Impactor
Whitby Electrical
Aerosol Analyzer
Royco Optical
Particle Counter
Diameter
Interval
Vim
0.05
0.1
0.2
0.4
0.8
1.6
3.2
6.4
Total
Loading
Solids
Loading
g/m3
0.019
0.027
0.013
0.050
0.088
0.070
0.013
0.28
Diameter Solids Diameter Solids
Interval Loading Interval Loading
ym g/m3 ym g/m3
0.015
0.020
0.06
0.110
0.1
0.408
0.15
1.31
0.3 0.4
0.38 0.048
0.6 0.8
0.053
1.5
0.024
3.0
2.23 0.125
49
-------
TABLE 17.
Approximate
Diameter
Range
IMPACTOR DATA (diameter range based
on a density of 2 g/cm3)
Wt% in Diameter Range
Celesco
Impactor
Lundgren
Impactor
0.05-0.4
0.4-.8
0.8-1.6
1.6-6.4
>6.4
21
18
31
29
0
88
4
3
4
0.5
60
54
12
5G
24
18
12
6
0
~
0.01
0.10
DIAMETER,
1.00
10.00
Figure 19. The mass distribution of the fly ash simulation aerosol
measured with a Celesco cascade impactor assuming a particle density
of 2 g/m3. (Courtesy of Dr. R. L. Chuan, California Measurements
Inc. Sierre Madre, CA. Note how this mass distribution compares
with the volume distribution shown in Figs. 17 and 18.
50
-------
cc
UJ
<
Q
10.0
8.0
6.0
4.0
2.0
1.0
0.8
0.6
0.4
0.2
0.1
0.08
0.06
0.04
0.02
0.01
I
0.01 0.1 1 10 30 50 70 90 99
CUMULATIVE PERCENT, STATED MICRON SIZE
99.9 99.99
Figure 20. The coal fly ash simulation aerosol size distribution
as a function of particle size as determined with a cascade impactor
(mass distribution) and with the EAA and Royco (volume distribution).
51
-------
Particle Resistivity
The particle resistivity has a large effect on the collection efficiency
of electrostatic precipitators. Electrostatic precipitators are installed in
many coal burning plants. Thus, it is important to have coal fly ash simula-
tion aerosols with various resistivities to conduct all of the necessary
precipitator tests. To accomplish this, lithium salts were added to the
basic solution (described in Section 4) used to generate the coal fly ash
simulation aerosol. LiNOs and LiClOi+OHaO are soluble in the solution; on
decomposing in the flame they yield Li2CO3 and LiCl.
Table 18 shows the particle resistivity of the simulation aerosol with
0.43 grams of LiaCOa per gram of simulation aerosol (SiOa, Fe203, Fe30i,
and Al203»2H20) and also another with 0.16 g of LiCl per gram of simulation
aerosol.
The resistivities were measured either with samples taken from the flame
holder above the flame cone or with samples collected by placing an aluminum
screen at the aerosol exit at the top of the stack. Only a small amount of
aerosol particles collected on the screen (about 2 grams). With such small
quantities, standard procedures for measuring resistivities could not be
used. The resistivities shown in Table 18 do, however, show that even with
small samples at room temperature and low humidity there is a difference in
the particle resistivities when lithium is present in the particles.
A large amount of lithium carbonate (0.43 g/g) was used in the simula-
tion aerosol to measure an extreme case. Smaller amounts of lithium carbonate
should be sufficient to adequately alter the particle conductivity. Figures
21 and 22 show the size and volume % distributions of the coal fly ash
simulation aerosol with lithium carbonate. Both are close to the measured
distributions of the fly ash simulation aerosol without added lithium. The
generation rate (as particle volume per m3 of fume) increased from 0.25
cm3/m3 without lithium to 0.39 cm3/m3 with lithium. With lithium added, the
Table 18. Particle Resistivity* of Coal Fly Ash
Simulation Aerosol (T=21°C, RH=27%)
Aerosol Collection Point Measured Resistivity*
(ohm cm)
Simulation aerosol Bottom of stack 1 x 1013
without lithium Top of stack 2 x 1012
With 0.43g Li2C03 added per g
simulation aerosol Bottom of stack 5 x 106
With 0.16g LiCl added per g Bottom of stack 2 x 108
simulation aerosol Top of stack 2 x 1011
*These resistivities were measured with small samples of material at room
temperature and may not be comparable to standard resistivity measurements.
52
-------
Ul
g
2
nIO
10s
106
105
0.01
o
DIAMETER, (im
Figure 21. The size distribution of
the coal fly ash simulation aerosol
with LiaCOs added to increase the
particle conductivity. The size
distribution is similar to that of
the simulation aerosol; adding the
Li2C03 had a small effect.
60
54
48
42
36
30
24
18
12
0.01
0.10
DIAMETER,
1.00
10.00
Figure 22. The volume distribution determined from the EAA
and Royco measurements of the coal fly ash simulation aerosol
with Li2C03 added.
-------
Royco counts were essentially unchanged. There was a reduction in the particle
concentration measured by the EAA (compare Figures 15 or 16 with 21). Thus,
there was) better agreement between the Royco and EAA measurements with added
lithium.
Sulfur Dioxide
CS2 is compatible with the basic solution. Immediately after mixing
there is a slight reaction with a few bubbles formed but it does not appear
to be a problem. The solution with €82 should be usable for a few hours
after mixing.
Wher. a solution of CS2 in ethanol (38 g CS2 per liter of ethanol) was
burned there were some small particles formed. No particles larger than
0.06 ym were detected (see Table 12). After combustion more than 99.8% of
the sulfur appeared to exist as S02 gas. The remainder (those particles
detected) was probably
Particle Size Distributions of the Components of the Simulation Aerosol
Since the simulation aerosol is generated by burning the basic solution
of 3 solutes [A1(N03)3«9H20, Fe(N03) 3 '9H20 and (CH3) 2Si(OC2H5) 2] , in ethanol,
simpler aerosols can be made by burning solutions with only one or two solutes.
Solutions of all the seven possible combinations (solvent plus 1, 2 or 3
solutes) were burned and the aerosol particle size distributions measured.
The weight (per liter of ethanol) of each solute present remained the same as
in the basic solution described in Section 4. Figures 15-16 and Appendix B
show the size distributions of these aerosols. By laying one figure over
another, it was easy to compare the measured size distribution of one aerosol
with another. The results of the comparisons follow:
1. The contribution of Fe203-Fe30u particles to the coal fly ash simu-
lation aerosol particle counts was very little; burning the basic solution or
the basic solution without Fe(N03) 3«9H20 resulted in about the same particle
counts. This was in agreement with the generation rates; 0.016 cm3/s as
Fe203-Fe3Cit generated vs. 0.16 cm3/s of the other oxides and also in agreement
with the cyclotron excitation analysis results (Table 15) where only 5% of
the aerosol volume was
2. The (CH3)2Si(OC2Hs)2 decomposed to form an aerosol which the electri
cal aerosol analyzer could not measure satisfactorily. Most of the aerosol
particle mass was composed of particles too small to be measured by the
optical particle counter.
3. When Si02 was present in any of the aerosols it completely dominated
the particle counts by the electrical aerosol analyzer (EAA); these counts
changed little regardless of what else was in the aerosol.
-------
ELECTRIC ARC FURNACE FUME SIMULATION AEROSOLS
Tables 19 and 20 show several of the characteristics of two electric arc
furnace (EAF) fume simulation aerosols. EAF aerosol A was generated using
the 0.25 m diameter stack without a reducer at the top. EAF aerosol B was
generated using the 0.25-m stack with a 0.15-m reducer at the top. The
reducer lowers the gas flow thus increasing the aerosol solids loading and
exit temperature.
Figures 23-26 show the size and volume distribution of electric arc
furnace simulation aerosols A and B. Figure 27 shows the EAA and Royco
particle counts of aerosol A plotted as AN/AlogD vs. D. For both aerosol A
and B the agreement between the EAA and Royco counts is good. The particle
loadings of A and B determined from the EAA and Royco counts agree with the
generation rates (see Table 9); the particle loading ratio, counted/generated,
was 0.40 and 0.26 for aerosol A and B, respectively. Aerosol B with a higher
solids loading has a larger median volume diameter, 0.2 vs. 0.08 ym for A.
Figure 28 shows N(D) vs. D for both aerosol A and B. There is not a great
deal of difference in these size distributions. Aerosol A has a higher
concentration of particles with D<0.13 ym and a lower concentration of
particles with D>0.13 ym. Other details of the particle counts are given in
Tables 9, 10, 12 and Appendix C.
TABLE 19. CHARACTERISTICS OF ELECTRIC ARC FURNACE
SIMULATION AEROSOL A
Stack: Diameter, 0.25 m; Height, 3.5 m
Fume Exit Temperature: 762°K (489°C)
Fume Exit Velocity: 7.6 m/s
Fume Flow Rate: 0.38 m3/s; 0.16 kg/s
Particle Loading: 0.65 g/m3 (1.8 g/m3 @ STP)
Particle Volume Median Diameter: 0.08 ym*
Condensation Nuclei: 3x10 8 par tides/cm3
Nominal Chemical Composition: 50% FeaOa,
20% Cr203,
15% MgO,
15% MnO
Nominal Density (Handbook values): 4.8 g/cm
Particle Color: Brownish black
*Median of those particles sized and counted
55
-------
TABLE 20. CHARACTERISTICS OF ELECTRIC ARC FURNACE
SIMULATION AEROSOL B
Stack: Exit diameter 0.15 m; Height 3.5 m
Fume Exit Temperature: 931°K (658°C)
Fume Exit Velocity: 12.5 m/s
Fume Flow Rate: 0.21 m3/s, 0.073 kg/s
Particle Loading: 1.2 g/m3 (4.0 g/m3 @ STP)
Particle Volume Median Diameter: 0.2 ym*
Condensation Nuclei 2.4x108 particles/cm3
Nominal Chemical Composition: 50% Fe203,
20% Cr203,
15% MgO,
15% MnO
Nominal Density (Handbook values): 4.8 g/cm3
Particle Color: Brownish black
*Median of those particles sized and counted
56
-------
10'
io1
10"
IO11
IO9
IO9 -
10'
10"
IO6
106
IO5
10*
IO3
103
IO2
10*
10'
1.00
DIAMETER, urn
0.10
DIAMETER, urn
Figure 23. Electric arc furnace
simulation aerosol A (0.65 g/m3)
size distribution from measure-
ments with the EAA and Royco.
Figure 24. Electric arc furnace
simulation aerosol B (2.1 g/m3) size
distribution from measurements with
the EAA and Royco.
57
-------
eo
!4
0.01.
0.10
DIAMETER, >un
1.00
10.00
Figure 25. The volume distribution determined from EAA and Royco
measurements of electric arc furnace simulation aerosol A
(0.65 g/m3).
60
54
413
42
33
s?
LU
I 3°
O
> 24
18
12
6
0
0.01
0.10
DIAMETER,
1.00
10.00
Figure 26. The volume distribution determined from EAA and Royco
measurements of electric arc furnace simulation aerosol B (1.2 g/m3)
58
-------
109
108
107
I 106
Z
•-1
105
103
102
0.01
EAA COUNTS
1.00
10.00
101'
109
10°
106
10"
103
0.01
0.10
1.00
10.00
DIAMETER, /am
DIAMETER, pm
Figure 27. Electric arc furnace
simulation aerosol A (0.65 g/m3)
size distribution, plotted as
AN/AlogD vs. D.
Figure 28. The measured size distri-
butions of the electric arc furnace
simulation aerosols A and B are
nearly the same.
BASIC OXYGEN FURANCE FUME SIMULATION AEROSOLS
Tables 21 and 22 show several of the characteristics of two basic oxygen
furnace fume simulation aerosols. Simulation aerosol AA (Table 21) has a
very high particle loading (6 g/m3 hot, 16.8 g/m3 @ STP) of red Fe203;
95% of the particle volume was contained in particles with diameters between
0.15 and 3 ym. Simulation aerosol AA was generated from a solution of
Fe(CO)5 in acetone (see Section 4). Simulation aerosol BB (Table 22) was
primarily FeaOi, with some FeaOs and Ca(OH)a*xH20; it had a large portion of
the particle mass in particles with diameters greater than 10 ym. It was
generated from a solution of Fe(NOs)3*9HaO and Ca(N03)2*4H20 in ethanol
(see Section 4). Lowering the concentration of solutes will reduce the
particle loading but will also reduce the portion of large particles.
59
-------
TABLE 21. CHARACTERISTICS OF BASIC OXYGEN FURNACE
FUME SIMULATION AEROSOL AA
Stack: Exit diameter, 0.15 m; Height, 3.5 m
Fume Exit Temperature: 764°K (491°C)
Fume Exit Velocity: 10 m/s
Fume Flow Rate: 0.17 m3/s; 0.073 kg/s
Particle loading: 6 g/m3 (16.8 g/m3 @ STP)
Particle Volume Median Diameter: 1.05 ym
Condensation Nuclei: 3.9 - 29xl08 particles/cm3
Nominal Chemical Composition: 100% FeaOa
Nominal Density (Handbook values): 5.12 g/cm3
Particle Color: Red
TABLE 22. CHARACTERISTICS OF BASIC OXYGEN FURNACE
FUME SIMULATION AEROSOL BB
Stack: Exit diameter, 0.15 m; Height, 3.5 m
Fume Exit Temperature: 913°K (640°C)
Fume Exit Velocity: 12.7 m/s
Fume Flow Rate: 0.21 m3/s; .076 kg/s
Particle Loading: 1.8 g/m3 (6.0 g/m3 @ STP)
Particle Volume Median Diameter: 0.3 ym*
Nominal Chemical Composition: 95% Fe203, 5% CaO
Nominal Density (Handbook values): 4.8 g/cm3
Particle Color: Black
Magnetic Properties: Highly magnetic
*Median of only those particles sized and counted with the EAA and Royco
60
-------
Figures 29-31 show the size and volume % distribution of aerosol AA.
The measurements were only approximate because only one reading was taken
for each size range. The instruments were then shut off to prevent damage.
Even with the 38,700 dilution factor too many particles entered the sample
train and most surfaces of the instruments in contact with the diluted
aerosol were left with a coating of red FeaOa. The particle counts of sizes
between 0.6 and 1.2 um were estimated because the Royco was saturated. The
Royco counted high background counts (e.g., 40 particles/cm3 with D>2 ym)
for several hours after the test. The EAA appeared to operate satisfactorily
until this experiment was completed; however, it would not function properly
for several days thereafter. Other details of the particle counts of aerosol
AA are given in Tables 9, 10, 12 and Appendix C.
Figures 12a and 12b show the volume distribution of aerosol BB.
,10 .
10
10s
108
10'
106
105
10"
103
102
10'
-EAA COUNTS
0.10
10.00
DIAMETER, um
Figure 29. Basic oxygen furnace
simulation aerosol AA (FeaOs 6 g/m3)
size distribution plotted as AN/A
logD vs. D.
n'O
109
108
10'
106
104
103
102
0.10
1.00
10.00
DIAMETER, (im
Figure 30. Basic oxygen furnace
simulation aerosol AA (FeaOa 6 g/m3)
size distribution from measurements
with the EAA and Royco.
61
-------
a?
UJ
60
54
413
42
36
30
24
18
12
C
0
0.01
0.10
DIAMETER,
1.00
10.00
Figure 31. The volume distribution determined from EAA and Royco
measurements of the basix oxygen furnace simulation aerosol AA.
Ninety-five percent of the particle volume was contained in parti-
cles with diameters between 0.15 and 3 ym.
ZINC SMELTER FUME SIMULATION AEROSOL
Tablo 23 shows several of the characteristics of the ZnO(0.94 g/m3)
aerosol described also in Tables 9, 10, 12 and Appendix C. This aerosol can
be used to simulate zinc smelter effluents. It was generated by burning a
solution containing 159 g of (73% pure) zinc nitrate per liter of methanol.
Ethanol can be used in place of methanol; with ethanol the exit fume temperature
rises. The basic solution for generating simulated zinc smelter effluent
shown in t:he Section 4 can also be used. Because of the toxicity of CdO,
this solution was not used in particle size measurement experiments. ZnO
(1.3 g/m ) fumes generated by burning solutions consisting of 234 g of (73%
pure) zinc, nitrate per liter of methanol also had volume median diameters of
0.2 ym. Figures 32-34 show the size and volume % distribution of the ZnO(0.94
g/m3) fume.
62
-------
TABLE 23. CHARACTERISTICS OF THE ZINC SMELTER
FUME SIMULATION AEROSOL
Stack: Diameter, 0.25 m; Height, 3.5 m
Fume Exit Temperature: 579°K (306°C)
Fume Exit Velocity: 5.8 m/s
Fume Flow Rate: 0.29 m3/s; 0.16 kg/s
Particle Loading: 0.94 g/m3 (2 g/m3 <§ STP)
Particle Volume Median Diameter: 0.2 ym
Condensation Nuclei: 1.4-1.7xl08 particles/cm3
Nominal Density (Handbook values): 5.47 g/cm3
Particle Color: white
10a -
108
106
10=
10J
10*
10'
0,01
1.00
DIAMETER, l
10"
107
I 106
105
10"
103
102
10.00
Figure 32. ZnO (0.94 g/m3) aerosol
size distribution from measurements
with the EAA and Royco.
10'
0.01
-EAA COUNTS
-ROYCO
COUNTS
I
0.10 1.00
DIAMETER, >im
Figure 33. ZnO (0.94 g/m3) aerosol
size distribution plotted as AN/AlogD.
63
-------
60
54
48
42
30 -
I!4
18
12
0.01.
0.10
DIAMETER, urn
1.00
10.00
Figure 34. The volume distribution determined from EAA and
Royco measurements of the ZnO (0.94 g/m3) aerosol.
64
-------
SECTION 7
COST OF RAW MATERIALS
Table 24 shows the materials cost for generating 45 kg (100 pounds) of
the various simulation aerosol particulates. Forty-five kg is a large
amount of aerosol. (i.e., The coal fly ash simulation aerosol has a particle
loading of 0.66 g/m3; therefore, 68,200 m3 (2.4xl06 ft3) of aerosol would
have to be generated to contain 45 kg of particles.)
The materials cost for a simulation aerosol for the zinc smelter is
much less if a pure ZnO aerosol is used. It would cost $1112 to generate 45
kg of the pure ZnO (0.94 g/m3) aerosol shown in Table 23 versus $2512 to
generate 45 kg of the zinc smelter aerosol shown in Table 24.
65
-------
TABLE 2
-------
SECTION 8
DISCUSSION OF RESULTS
The program objective of developing reproducible stable sources of
well-characterized solid particulate, fine particle aerosols for laboratory
pilot plant tests of pollution abatement equipment was accomplished. Methods
were developed for generating fine particle aerosols which have chemical
compositions typical of four industrial effluents. The S02 content of the
carrier gas can be regulated at any desired concentration. Particle resistiv-
ities can be lowered by adding lithium salts. The volume median diameter of
the particles can be varied from 0.06 to 5 ym by adjusting solution composi-
tions, concentrations or the stack exit diameter. Aerosol particle loadings
can be varied from less than 0.01 g/m3 to 6 g/m3. Magnetic, non-magnetic,
sticky, not sticky, highly resistive, water soluble, hydroscopic and hydro-
phobic aerosols were generated.
All twelve of the metal nitrates and other compounds (excluding
in H20 and CHsOH which are probably incompatible with (CH3)Si(OC2H5) 2)
used in these programs are believed to be compatible in solution with each
other. Thus, aerosols consisting of several thousand different combinations
of chemical compounds can be generated. An infinite number of weight ratios
can be formulated. There are also several other metal nitrates and chlorides
which are soluble in ethanol and/or methanol. They could be used to generate
aerosols of other metal oxides and chlorides. The most useful aerosols,
however, will probably be those comprised of a single compound.
When an aerosol is generated such that all of its particles consist of
a single compound and crystal type (e.g., face centered cubic MgO) , then all
of the particles will have most, if not all, of the same properties such as
density, resistivity, permitivity, etc. It should be possible to determine
any remaining size dependent properties using the current knowledge of solid
state physics. By using several different simple aerosols, properly chosen
to have a wide range of properties, it should be possible to obtain data which
will lead to the determination of basic theories, as opposed to empirical
findings, of the important parameters of specific pollution abatement equip-
ment. This knowledge could then be used in models to predict which particle
sizes of particular compounds in industrial effluents would not be collected
by various pollution abatement equipment. Tests conducted with simple aerosols
of a single compound and crystal type will simplify the analysis and will
yield less ambiguous results. When using a simple aerosol, the analytical
costs will be much less than when using a complex aerosol; and the confidence
of the analysis will be higher.
67
-------
Complex aerosols consisting of several chemical compounds usually have
particles with differing compositions and properties. These aerosols probably
would not be as useful in determining basic pollution abatement equipment
parameters because of the analytical difficulties. The characterization of
the coal fly ash simulation aerosol showed that the average chemical composition
of the particles was a function of size (see Table 15, cyclotron excitation
analysis) and that in a given size range the chemical composition varied
from particle to particle (see the X-ray Non-dispersive Analysis Section).
These are properties typical of real fly ash and other industrial effluents
(see Section 3). Based on these findings one would expect that the particles
might have varying properties depending on their composition. They also may
have surface compositions which are different from the bulk particle composi-
tions. It would be difficult to determine all of the various aerosol particle
properties and to determine whether each is due to the bulk or surface
compositions. For example, a 0.2 ym Fea03 particle with hundreds of 0.01 ]M
SiOa particles coating its surface would probably have a resistance comparable
to Si02 not Fe203.
Use of single compounds is safer than using complex compounds. Toxic
materials! could be avoided. Once the basic principles are better known, the
important: properties of materials such as CdO probably can be simulated by
other non-toxic materials and the abatement equipment tests could be performed
with a non-toxic material.
The make-up of the surface of a particle determines many of the physical
properties of the particle. The flame method used to generate particles
with freeh surfaces is superior to resuspension of old materials. It closely
duplicates the actual effluent sources. As particles age, the surface
properties change. It is also unlikely that submicron sized long chain
aggregates can be collected then resuspended so that they duplicate the long
chain aggregates contained in industrial effluents.
Many trace elements condense onto the surface of effluent particles
(Davison, 1974; Pueschel 1976). The particle's surface properties will
often be a function of this surface coating, not the bulk composition. In
order to adequately simulate these effluents, it may be necessary to generate
simulation aerosols with particle composition different from the bulk effluent
particle compositions.
The plots of size distributions and Tables 11 and 12 show that the EAA
and Royco particle counts are in good numerical agreement for some aerosols
and in poor agreement for others. The instrument particle counts given in
the Results Section were not, however, corrected for specific aerosols
properties (e.g., optical properties). The Royco Model 200 optical particle
counter has a right angle optical system and the particle light scattering
to the detector is a function of the particle shape and refractive index.
Table 25 shows the optical properties of several of the generated aerosol
materials. Cooke and Kerker (1975) showed the strong dependence of the
Royco Modiil 200 optical particle counter response to both the refractive
index and adsorption coefficient of a particle.
68
-------
TABLE 25. OPTICAL PROPERTIES
Compound
Si02
MgO
ZnO
Fe304
Mn304
Cr203
Fe203
Refractive
Index
(A=0.589 ym)
1.5
1.74
2.01
2.42
2.15-2.46
2.55
3.01-3.46
Adsorption
Coefficient
~0
~o
~0
0.55-0.57
0.5-l(Est.)*
-
0.97-1.07
Particle
Count ratio
EAA/Royco
0.32 ym). The India Ink particles were shown to be spheres with rough
surfaces. Gebhart (1976) states that for particles smaller than the wave-
length of light, scattering becomes more and more a volume effect and depends
only on the optical properties. As long as light scattering is a volume
effect, the scattering cross-section of a particle becomes mainly a function
of its volume and less of its shape. Consequently, a light-scattering
device used for particles smaller than the wavelength of the light measures
the equivalent volume diameter of a non-spherical particle.
Since the Royco counts of the background laboratory air were normally
lower than the EAA counts by a factor of 7.7 and since the particles in the
laboratory air should have typical optical properties, there appears to be a
systemmatic error in at least one of the instruments. The EAA was calibrated
with a Model 3030 EAA and the particle counts of both instruments in the
size range 0.3
-------
good, tht! refractive index must have caused the Royco to "over size" and the
resulting agreement was coincidental.
The poorest agreement between the EAA and Royco was with SiOa particles
which were transparent and with the MgO particles which are white. Since
the concentration of particles decreases roughly in proportion to the cube
of the diLameter, under or over sizing has a large effect on the number of
particles counted in a size range. For SiOa aerosols the particle concentration
in the s:lze range 0.3
-------
Table 10 shows aerosol particle concentrations measured by the Environment/
One condensation nuclei counter (CNC) and by the EAA. Dividing these concentra-
tions by the dilution (6,900, for three stage and 38,700 for non-standard
three stage) converts the aerosol concentrations to the instrument counts of
the diluted aerosols. Liu and Pui (1974) compared the Environment/One and
EAA at several NaCl particle concentrations. They showed that the Environment/
One CNC had a linear response at concentration levels up to 52,000 particles/
cm3 (indicated concentration) but the indicated concentration was lower than
the true concentration by a factor of 2.5. At higher concentrations the
instrument response was nonlinear; at 140,000 particles/cm3 indicated, the
indicated concentration was lower than the true concentration (600,000) by a
factor of 4.3. These findings of Liu and Pui were used to correct the CNC
measurements shown in Table 10. After correction, the EAA counts (see Table
10) of the total particle concentrations were higher than the corrected CNC
measurements by a factor of 2.45. For the HaSOit'xHaO droplet aerosol the
EAA counts of the particle concentrations were lower than the corrected CNC
measurements by a factor of 2.1.
Tables 26-29 summarize comparisons of the effluents and the simulants.
The zinc smelter effluent can also be simulated by the simulation aerosol
generated from the basic zinc smelter solution shown in Section 4. This
aerosol consists of ZnO and CdO with some V02, V^Of, V20s, ZnSOi,, CdSOi+,
and 2CdO'V205.
TABLE 26. COMPARISON OF COAL FLY ASH EFFLUENTS WITH SIMULANT
Temperature C
Particle Loading (g/m3)
Mass or Volume Median Diameter (ym)
Particle Density (g/cm3)
Bulk Density
Electrical Resistivity (ohm cm)
Particle Composition
Effluents
121-157
2.3-12.8
-15 ym
0.6-3
0.5-0.8
108-1013
Si02
Fe203
A1203
CaO
MgO
Na20
Simulant
177
1.1
0.2-1
2.6
0.3-0.9
5xl06-2xl012
Si02
Fe203
Al203-xH20
71
-------
TABLE 27. COMPARISON OF ELECTRIC ARC FURNACE EFFLUENTS WITH SIMULANTS
Temperature (°C)
Particle Loading (g/cm )
Mass or Volume Median Diameter (ym)
•3
Particle Density (g/cm )
3
Electrical Properties (ohm cm )
Particle Composition
Effluents
100-1650
0.2-5
0.3-5.4
3.8-3.9
6xl05-7xl013
Fe203, FeO,
Cr203, Si02,
A1203, CaO,
MgO, MnO,
ZnO, CuO,
NiO, PbO,
C, S, P,
Alkalies
Simulants
489-658
1.8-4
0.08-0.2
4.8
Not Measured
Fe2°3
Fe3°4
Cr203
MgO
MnO
TABLE 5.8. COMPARISON OF BASIC OXYGEN FURNACE EFFLUENTS WITH SIMULANTS
Effluents Simulants
Temperature (°C) 290-1650 491-640
Particle Loading (g/cm3) 4.6-23 6-16.8
Mass or Volume Median Diameter (ym) 0.095 0.3-1.05
Particle Density (g/cm3) 3.44
Particle Composition Fe°
Particle Density (g/cm3) 3.44 4.8-5.1
o 9
-J £•
FeO or
MnO Fe30
Si0 Fe0
Ca(OH)2-xH20
CaO
MgO
72
-------
TABLE 29. COMPARISON OF ZINC ROASTER AND SINTERING MACHINE
EFFLUENTS WITH SIMULANT
Temperature ( C)
Particle Loading (g/cm )
Mass or Volume Median Diameter (ym)
Particle Density
Particle Composition
Effluents
PbO
CdO
Simulants
160-482°C
0.9-150
0.4-15
-
ZnO
306
2
0.2
5.47
ZnO
These simulation aerosols appear to have characteristics similar to those of the
effluents and should suffice for many pollution abatement equipment tests.
73
-------
REFERENCES
1. Brnkhnova, I. T. Environmental Hazards of Metals, Consultants Bureau,
New York, 1975, 277 pp.
2. Carroz, J. W. and R. C. Noles. A new ground based solution burner for
cloud seeding and production of atmospheric tracer materials. Third
Conference on Weather Modification, Am. Met. Soc., June 26-29, 1972,
pp. 37-39.
3. Gavin, D. C., W. A. Klemm and G. Burnet. Analytical methods for
characterization of fly ash. AEC Ames Laboratory, Ames, Iowa, 1974,
19 pp., IS-M-14.
4. Chuan, R. L. Rapid measurement of particulate size distribution in the
atmosphere. Fine Particles edited by B.Y.H. Liu, Academic Press, N.Y.,
1976, pp. 763-775.
5. Cooke, D. D. and Milton Kerker. Response calculations for light-scattering
aerosol particle counters. Applied Optics, 14(3), March 1975, pp. 734-739.
6. Davison, R. L., D. F. S. Natusch, J. R. Wallace, C. A. Evans Jr. Trace
elements in fly ash. Environ. Sci. Tech., 8(13), Dec. 1974, pp. 1107-1113.
7. Drehmel, D. C. Primary fine particle control technology. 69th Annual
Meeting of the Air Pollution Control Association, Portland, Oregon,
June 27-July 1 1976, 14 pp.
8. Drehmel, D. C. and D. E. Black. Particulate control in energy processes:
a status report. J. Air Pollution Control Assoc., 26(12), December 1976,
pp. 1141-1143.
9. Gebhart, J. J., Heyder, C. Roth, W. Stahlhofen. Optical aerosol size
spectrometry below and above the wavelength of light - a comparison.
Fine Particles edited by B.Y.H. Liu, Academic Press, N.Y., 1976, pp. 793-815.
10. Harvey, R. D. Petrographic and mineralogical characteristics of carbonate
rocks related to sulfur dioxide sorption in flue gases. Illinois State
Geological Survey, July 1971, 93 pp., PB 206-487.
11. Harris, D. B.., D. C. Drehmel. Fractional efficiency of metal fume
control as determined by Brink impactor. Presented at 66th Annual Meeting
APCA Chicago, 111., June 24-28, 1973, 26 pp.
74
-------
12. Liu, B.T.H. and D.Y.H. Pui. A submicron aerosol standard and the primary,
absolute calibration of the condensation nuclei counter. J. of Colloid
and Interface Science, 47 (1), April 1974, pp. 155-171.
13. Luke, W. I. Nature and distribution of particles of various sizes in
fly ash. T. R. 6-583, Corps of Engineers, Vicksburg, Mississippi,
Nov. 1961, 21 pp, AD-731-654.
14. Pueschel, R. F. Aerosol formation during coal combustion: condensation
of sulfates and chlorides on fly ash. Geophysical Research Letters,
3 (11),November 1976, pp. 651-653.
15. Sem, G. H. Submicron particle sizing experience on a smoke stack using
the electrical aerosol size analyzer. Presented at the Seminar on
In-Stack Particle Sizing for Particulate Control Device Evaluations,
Environmental Protection Agency, National Environmental Research Center,
Research Triangle Park, N.C., December 3-4, 1975.
16. Shen, T. T., et^ al. Characterization of differences between oil-fixed
and coal-fired power plant emissions. To be presented at the Inter-
national Clean Air Congress, Tokyo, Japan, May 1977.
17. Vandegrift, A. E., e_t al. Particulate pollutant system study. Vol. Ill -
Handbook of Emission Properties. Midwest Research Institute, 1 May 1971,
607 pp., PB-203-522.
18. Whitby, K. T. and R. A. Vomela. Response of single particle optical
counters to nonideal particles. Environ. Sci. Tech., 1 (10), October 1967,
pp. 801-814.
19. Yakowitz, H., M. H. Jacobs, and P. D. Hunneyball. Analysis of urban
particulates by means of combined electron microscopy and X-ray
microanalysis. Micron, 3: pp. 498-505, 1972.
20. Author unknown. Engineering and cost study of the Ferroalloy Industry.
EPA, Research Trinagle Park, North Carolina, EPA 450/2-74-008, May 1974,
pp. D20-D28.
75
-------
APPENDIX A
OPERATING SPECIFICS
OPERATING PROCEDURE
Tliis operating procedure is to be used with the generation apparatus
descritied in the text.
The procedure for safe operation of the generation apparatus consists of
two important points.
1. Do not use any system for pressurizing the pressure tank which will
cause the tank pressure to rise above 200 psi, also keep a pressure relief
valve on the tank in good operating order.
2. Do not expose people to dangerous levels of toxic chemicals. Mix,
spray and burn the chemicals described in the final report only with adequate
protection.
The generation of reproducible aerosols requires the use of "good" solu-
tions. With one possible exception, the solutions described in Section 4 of
the final report should be mixed just prior to use. They will degrade within
several hours after mixing. The stability of iron pentacarbonyl in acetone
was not measured. However, all of the nitrates, except for ferric nitrate,
are stable for months when dissolved in ethanol. Ferric nitrate in ethanol
forms a significant amount of insoluble material within 24 hours. When chromic
and ferric nitrate are in solution together an insoluble compound forms
within several hours. Many of the other nitrate compounds named in Section 4
of the final report are probably compatible and stable for long periods of
time whem dissolved in ethanol. Until the stability of a particular solution
has been tested, it should be used within a few hours after mixing.
After generating aerosol, the hardware should be washed several times with
water, the nozzle should be disassembled and then washed. Some of the nitrates
are very corrosive. Other operating parameters and a description of the
apparatus are given in Section 4 of the final report. They include using
200 psi solution pressure and a WDB 4.0, 30° Delavan nozzle.
DILUTER DESCRIPTION
Parts List (All dimensions in English Units to aid in construction)
1, Fume Intake Tube: 1/2 inch stainless steel tubing with 7/16 inch ID.
76
-------
2. Porous Metal Tube: 2 Inch OD, 9 inches long, 1/8 inch wall thickness,
grade F-30 sintered brass (Pacific Sintered Metals Co., 16120 So. Figueroa St.,
Gardena, Calif. 90247).
3. Brass T: 10 inches long, 3 1/2 inch OD. brass tube. Dilution air
intake tube has a 4 inch long, 2 1/8 inch OD copper arm 2 inches from fume
intake end. Copper arm is made from 2 1/8 inch OD, 1 31/32 inch ID, Standard
"2 inch" copper pipe.
4. Phenolic End Piece: 2 inches thick, OD is machined to fit ID of brass
T. Also machined to hold fume intake probe and porous metal tube (see Figure 5).
5. Brass End Piece: 1 1/4 inches thick, OD is machined to fit inside
brass T and to hold porous metal tube and copper pipe.
6. Copper pipe: 24 inches long, copper pipe (Std. "2 inch pipe", 2 1/8 inch
OD, 1 31/32 inch ID).
7. Mixer: 2(ea) thin (0.030 inch) brass disks. OD of upstream disk =
2 1/16 inch; OD of downstream disk = 1 15/16 inch; each disk has 6(ea) 1/2 inch
holes punched in it (see Figure 2E). Disks held together with 2 3/8 inch long
piece of 1/4 inch copper tubing.
8. Elbow: Std. copper 90 elbow with standard "1/2 inch" copper pipe
soldered in so to be coaxial with the 2 inch copper pipe (see Figure 2 G ).
9. Other Metal Parts: Standard 1/2 inch and 2 inch pipe and 90° elbows
are used to complete the set up. Standard 1/2 inch copper pipe has a .52 inch
ID. All fittings are fastened together with duct tape. Standard copper sweat
fitting reducers are used at B and G for "orifices". At C and E paper washers
are used.
10. Fan: Model HH33 Quick-Air blower-suction unit operated at low speed.
Inlet blocked with 31/32 inch ID orifice, outlet blocked with 15/16 inch ID
orifice to reduce fan suction (Clements Mfg. Co., 6650 S. Narragansett Ave,
Chicago, 111. 60638).
Orifice Sizes (All dimensions in English Units to aid in construction)
Diluters 3(ea) (see Figure 4)
1. First Stage Diluter (see Figure 5)
(a) Fume intake (see Plate 4)
(b) Orifice ID @ B = 1/4 inch
(c) No orifice @ C, ID = 1 31/32 inches
(d) Orifice ID @ G = 3/8 inch
(e) Orifice ID @ I = 3/4 inch
77
-------
2, Second Stage Diluter (see Figure 5)
(a) Intake at G in Stage 1
(b) No orifice @ B, ID = 1/2 inch
(c) Orifice ID @ C = 1 1/4 inches
(d) Orifice ID @ G = 3/8 inch
(e) Orifice ID @ I = 1 inch
3. Third Stage Diluter (see Figure 5)
(a) Intake at G in stage 2
(b) No orifice @ B, ID = 1/2 inch
(c) Orifice ID @ C = 1 1/16 inches
(d) Orifice ID @ G = 3/8 inch
(e) No orifice @ I, ID = 1 31/32 inches
78
-------
APPENDIX B
AEROSOL SIZE DISTRIBUTIONS
The figures contained in this appendix show size distributions, not shown
in the text, of most of the aerosols which were generated and measured. The
figures are plotted as N(D) vs. D.
Refer to Tables 8 through 12 in text for additional information on these
aerosols.
79
-------
10'
0.10 100
DIAMETER. *jm
B-l. Si02(0.36g/m3)
106
I05
104
DIAMETER, ,,m
B-2. Si02 § Al263'2H20(0.62g/m3)
106
10=
104
B-3. Si02, Fe203
.51g/m3)
DIAMETEH. om
B-4. Fe203 S
used reduced pressure.
80
-------
10"
102
DIAMETER, urn
B-5. Al203'2H20,Fe203 & Fe30,,
(O.Alg/m3)
10'°
9
z
• 106
I03
B-6. Al203'2H20(0.26g/m3)
10'" -
B-7. MgO(0.054 & 0.49g/ra3)
B-8. ZnO(0.087 & 0.94g/m3)
81
-------
10"
ID4
10*
B-9. Fa203(0.071 & 0.56 g/m3)
105
103
10*
B-10. Mn3OU0.62g/m3)
B-U.
B-12. Ca(OH)2(0.29g/m3)
82
-------
APPENDIX C
PARTICLE CONCENTRATIONS AND VOLUME PERCENTS
This appendix shows the particle concentrations and the resulting
volume percents for 26 different aerosols. The diameter intervals are shown
at the top of each page, the aerosols are described by their particle composi-
tion followed in parenthesis by the particle loading (g/m3). The dilution
each aerosol underwent before measurement is given in Tables 11 and 12 in the
text. To convert aerosol concentrations back to the counts measured by the
instruments divide the concentrations shown in this appendix by 6900 for
standard three stage dilution and 256 for two stage dilution. For non-
standard three stage dilution divide the concentrations derived from the EAA
counts by 38,700 and the concentrations derived from the Royco counts by
65,700.
The following Whitby-Cantrell constants were used to convert the EAA
outputs to particle counts: The original data reduction constants given in
the instrument manual are also shown.
Diameter Constant Original Constant
.0075 1.5E5 9.1E5
.015 8.8E4 2.3E5
.03 5.9E4 6.2E4
.06 1.5E4 2.2E4
.1 1.2E4 1.2E4
.15 8.8E3 6.8E3
' 3.3E3 3.3E3
83
-------
PARTICLE CONCENTRATIONS AND VOLUME PERCENTS
.0075 .015 .03 .06 .1 .15 .3 .6
.3 .4 .5 .6
1.0 1.2 1.5 2.0 3.0 4.0 6.0
3.6E8 6.1E7 2.0E7 2.0E8 1.8E8 1.2E8 3.4E6
.03 .8 5 18 63 14
4.9E7 1.9E8 2.3E8 2.6E7 2.8E6 6.1E5 2.4E4
.1 4 40 27 12 13 4
Si02(0.36)
SiO.,(0.009)
Si02,Al 0 -2H 0(0.62) 3.0E7 1.8E8 2.2E8 2.5E8 1.6E8 5.4E6
.01 .5 4 18 59 16
SiO Fe 0 Fe 0 (0.51) 9.7E7 1.1E8 2.3E8 2.0E8 8.5E7 2.0E6
.1 .6 7 25 55 11
Coal Fly Ash! Simulation 0 1.1E8 2.4E8 2.3E8 2.1E8 1.3E8 4.7E6
Aerosol(0.66) .05 .8 5 18 57 17
Coal Fly Ash Simulation 0 2.9E8 4.7E7 1.8E8 2.3E8 1.2E8 3.8E6
Aerosol(0.66) .1 .2 4 20 54 14
Same as above with 1.1E8 9.7E6 2.1E8 2.2E8 1.3E8 5.3E7 1.7E6
.009 1.5 9 22 48 12
00 Li2C03(0.94)
5.9E8 3.3E8 1.6E7 1.8E6 6.1E5
11 49 14 6 11
Fe 0 ,Fe~0 (0.12) + 4.2E8 1.1E8 5.2E6 3.8E5 1.3E5 1.0E4
22 47 13 4 7 4
Al 0 -2H20,Fe 0 , 5.0E8 4.4E8 3.2E7 1.8E6 3. OE5 4.2E4
FeA(0.41) 3 22 10 3 3 2
A120 -2H20(0.26) 1.2E6 5.3E6 1.0E5 9.2E5 6.6E5 1.6E5
.02 .7 .1 3 11 20
S102(0.096)
MgO(0.62)
4.0E8 2.7E8 7.1E8 3.6E8 1.6E8 4.4E7 1.4E6
.3 5 16 28 40 10
(EAA saturated all stages <6)
•f Solution pressure reduced to 0.7 MPa (100 psi)
* MgO(0.62): AN=1.3E3, Vol2=16 (4
-------
PARTICLE CONCENTRATIONS AND VOLUME PERCENTS
.0075 .015 .03 .06 .1 .15- .3 .6 .3 .4
.5
.6
•8 1-0 1.2 1.5 2.0 3.0 4.0 8.0 10.0
MpO(0.054)
M£0(0.49)
ZnO(0.087)
ZnO(0.94)
4.5E7 1.5E8 3.2E8 1.3E8 5.9E7 3.2E7 1.6E6
.24 9 16 46 19
0 1.9E8 2.0E8 1.6E8 2.2E8 2.7E8 2.6E7
0 .03 .3 1 7 47 36
2.1E8 4.1E8 8.5E8 2.0E8 3.0E7 7.1E6 3.6E5
1 19 28 17 20 8
0 0 2.1E8 2.8E8 1.8E8 1.1E8 4.2E6
0 0 .86 16 50 16
2.5E4 2.0E4 1.4E4 1.1E4 4.2E3 2.1E3 3.6E3 5.4E2 4.6E2 3.1E2
.2 .3 .3 .5 . .5 .4 1 .41 2
2.4E5 2.1E5 1.5E5 9.8E4 6.6E4 3.7E4 1.9E4 8.4E3 4.0E3 4.4E3
.2 .4 .4 .6 .8 .9 .9 .8 1 3
1.9E4 1.3E4 1.2E4 6.5E3 6.1E3 2.8E3 1.6E3 8.8E2 3.1E2
.2 .4 .6 .6 1 1 1 1 1
1.2E6 4.9E5 1.5E5 7.1E4 5.7E4 2.4E4 7.7E3 1.0E4 4.9E3 4.1E3
2 1 .7 .7 1 .9 .5 2 2 5
Fe 0 (0.071)** 3.4E7 4.8E8 8.3E8 1.7E8 3.2E7 1.0E7 2.5E5
- J 1 20 24 18 31 6
5.9E3 1.6E3 8.8E2
.08 .05 .05
Fe,03(0.56)**
1.4E9 1.4E8 2.3E8 2.5E8 1.8E8 9.2E7 4.5E6
.1 .08 1 7 20 52 20
1.7E6 8.4E5 1.9E5 9,3E3
4 5 2. .2
oo
<•" Fe 0 (0.78)
^
1.5E10 3.7E9 2.1E9 3.5E8 3.0E8 4.4E8 2.3E7
.07 .1 .6 .6 2 15 6
3.6E8 1.2E8 2.8E8 1.2E7 1.4E6 1.8E6 2.3E5
.41 21 5 3 17 17
saturated
9.0E6 8.9E6 5.7E6 3.4E6 7.7E5 3.6E5 1.2E4 7.2E2
16 18 9 11 1 .3
1.5E5 1.0E5 4.7E4 2.6E4 1.2E4 5.0E3 2.5E3 1.0E3 2.3E2
698886653
Mn304(0.62)
3.7E8 6.5E8 2.8E8 2.1E7 4.3E6 4.1E6 5.5E5
.23 10 5 4 19 22
2.4E5 2.5E5 2.0E5 1.0E5 3.9E4 8.8E3 1.9E3 9.1E2 2.6E2
5 11 16 16 13 5 2 2 2
Cr203(0.69)
5.0E7 2.3E8 5.7E8 1.5E8 3.4E7 1.5E7 9.7E5
.124484
1.6E5 1.3E5 5.7E4 4.9E4 3.2E4 1.0E4 8.0E3 1.1E4 4.6E3 3.8E3 3.6E3 5.7E2
.4 .7 .5 1 1 1 1 3 4 9 36 23
Ca(OH) (0.29) 3.0E8 8.7E8 3.2E8 2.0E7 3.8E6 4.7E6 4.8E5
2 .02.4 1 .4 .3 2 2
9.3E4 1.4E5 1.0E5 8.5E4 8.3E4 5.1E4 4.1E4 2.7E4 1.7E4 8.9E3 *
.2 .5 .8 1 3 3 5 7 13 18
Ele. Arc Aerosol(0.65)1.1E8 9.0E8 4.8E8 2.8E7 3.7E6 1.6E6 1.4E5
.18 35 12 7 15 10
1.2E5 6.1E4 2.3E4 7.5E3 4.4E3 1.1E3 9.2E2 3.9E2 1.5E2
554231222
Ele. Arc Aerosol(1.2) 1.8E8 7.7E8 3.5E8 1.2E7 4.6E6 2.6E6 2.5E5
.02 6 22 5 7 20 16
1.9E5 1.2E5 5.1E4 1.6E4 6.0E3 2.1E3 1.3E3 1.5E2 5.4E2 1.5E2
6974323165
H.SO.-xH.O
24 2
2.5E9 4.0E7 1.8E7 nil
63 8 29
* Ca(QH)2(0.29): AN=7.5E3, Vol%=43(4
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-77-132
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Generation of Fumes Simulating Particulate Air
Pollutants
5. REPORT DATE
Julv 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
J.W. Carroz, F.K. Odencrantz, and W. G. Finnegan
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Department
Naval Weapons Center
China Lake, California 93555
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADM-031
11. CONTRACT/GRANT NO.
EPA Interagency Agreement
IAG-D5-0669
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Ofllce of Research and Development
Industrial Environmental Research Laboratory
Research. Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final: 1/75-4/77
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES IERL_RTp pr()ject officer for this report is Dennis C. Drehmel,
Mail Drop 61, 919/541-2925.
16. ABSTRACT Tne report describes techniques developed for generating large quantities
of reproducible, stable, inorganic, fine-particle aerosol fumes. These fumes simu-
lated particulate air pollutants emitted from power generation, basic oxygen furnaces,
electric arc furnaces, and zinc smelting. The aerosols were generated by burning
flammable solutions containing appropriate soluble compounds (e.g. , nitrates) of the
desired elements. In the flame, these compounds decomposed to oxides. Particle size
determinations were made using scanning and transmission electron microscope (SEM
and TEM) photographic analysis of captured particles, as well as Whitby and Royco
aerosol analyzers. The generated aerosol flow rates were as high as 42 cu m per min
'148 cfm); particle loadings were as high as 16. 8 g per cu m at STP. For most aerosols,
the aerosol particle and condensation nuclei concentrations were of the order of 10 to
:he 9th power particles per cu cm. The aerosol volume median diameters varied from
less than 0.015 to greater than 4. 7 micrometers and were primarily a function of the
solution ingredients. Methods were developed to vary the SO2 concentration and par-
icle resistivities.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution, Dust Control, Measurement
Particles, Combustion Products
riue Gases, Fly Ash, Oxides
inorganic Compounds, Fumes, Simulation
Slectric Power Generation
Basic Converters, Oxygen Blown Converter
Industry. Smelting
Air Pollution Control
Stationary Sources
Fine Particles
Condensation Nuclei
13B
21B
10A
11F
14B
07B
13H
3. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
97
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-' (9-73)
86
------- |