U.S. Environmental Protection Agency Industrial Environmental Research      EPA~600/7~78~01 3
Office of Research and Development Laboratory                  .     An^Q
                 Research Triangle Park, North Carolina 27711 rQoTU3T}f 1978
       CHARACTERIZATION
       AND GENERATION
       OF METAL AEROSOLS
      Interagency
      Energy-Environment
      Research and Development
      Program Report

-------
                  RESEARCH REPORTING SERIES


Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was  consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

    1. Environmental Health Effects Research

    2. Environmental Protection Technology

    3. Ecological Research

    4. Environmental Monitoring

    5. Socioeconomic Environmental Studies

    6. Scientific  and Technical Assessment Reports (STAR)

    7 Interagency Energy-Environment Research and Development

    8. "Special"  Reports

    9. Miscellaneous Reports

This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series.  Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data  and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects;  assessments of, and development of, control technologies for  energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.



                        EPA REVIEW  NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, 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 Informa-
tion Service, Springfield, Virginia 22161.

-------
                             EPA-600/7-78-013
                                February 1978
 CHARACTERIZATION
   AND GENERATION
 OF METAL AEROSOLS
               by

Neil Zimmerman, Dennis C. Drehmel, and James H. Abbott

       Environmental Protection Agency
      Office of Research and Development
    Industrial Environmental Research Laboratory
    Research Triangle Park, North Carolina 27711
        Program Element No. EHE623
            Prepared for

   U.S. ENVIRONMENTAL PROTECTION AGENCY
      Office of Research and Development
         Washington, D.C. 20460

-------
                                   Abstract

     The techniques of metal aerosol generation are reviewed in this
report to establish the state-of-the-art of the technology and to guide
future researchers.  Exposure to metal or metallic compound submicron
aerosols is widespread in both the industrial and general environments.
Research in areas of health effects, sampling instrumentation, and air
pollution control technology requires a reliable source of test aerosol.
Many mechanisms for the generation of metal aerosols are presented and
their applications, advantages, and disadvantages discussed.  Generation
methods can be on a continuous or a batch basis, with high or low concentrations
and generation rates, and with monodisperse or polydisperse size distributions.
The method chosen and the resulting aerosol depend on the requirements
of the specific research being conducted.  Metal aerosols generated by
exploding wires, electric arcs, heating to evaporation, combustion, and
dispersion are presented, with particular attention paid to particle
size characteristics.
                                        n

-------
                              Table of Contents

Section                                                           Page

Abstract                                                           ii
List of Tables                                                     iv
Conversions                                                        v
Terms and Abbreviations                                            v
Acknowledgements                                                   vi
1    Introduction                                                  1
1.1  Metal Aerosols from the Iron and Steel Industry               2
1.2  Metal Aerosols from the Nonferrous Metals Industry            4
13.  Metal Aerosols Characterized from Occupational Studies        6
1.4  Characteristics of Generated Metal Aerosols                   7
2    Exploding-Wire Generators                                     9
3    Electric Arcs                                                 13
3.1  Low  Intensity Arcs                                            15
3.2  High Intensity Arcs                                           18
4    Heating to Evaporation                                        21
4.1  Metal Oxides                                                  23
4.2  Metal Hal ides                                                 24
4.3  Metallic Aerosols in Argon                                    25
5    Combustion                                                    25
6    Dispersion                                                    28
6.1  Using Liquid Droplets—General                                28
6.2  Using Liquid Droplets—Nebulizers                             30
7    Summary                                                       31
8    References                                                    35

-------
                                   Tables
Number                                                               Page

   1           Summary of Particle Size Data for the                  3
               Iron and Steel Industry

   2           Summary of Particle Size Data for the                  5
               Nonferrous Metals Industry

   3           Summary of Particle Size Data from                    14
               Exploding-Wire Generators

   4           Summary of Particle Size Data for                     29
               Generators Using Combustion
                                     iv

-------
       Conversions
1 1
1 Ib
1 in.
1 ton
1 m /min
1 9
1 gr/ft3
1 ft3/min
= ICf4 ym
= 454.5 g
= 2.54 cm
= 907.2 kg
= 35.515 ft3/min
= 15.43 gr
= 2.29 g/m3
= 28.32 1/min
 Terms and Abbreviations





cL     =  aerodynamic diameter
 a


d   .   =  mass median diameter



PSD    =  particle size distribution



EM     =  electron microscope



K      =  coagulation coefficient



SEM    =  scanning electron microscope



TEM    =  transmission electron microscope

                                     o

n      =  concentration, particles/cm



a      =  geometric standard deviation
 3


CMD    =  count median diameter



MAD    =  median aerodynamic diameter

-------
                           Acknowledgments
     Discussions with Professor Benjamin Linsky of West Virginia University
and Dr. Parker Reist of the University of North Carolina at Chapel Hill
were very helpful and informative.
                                    VI

-------
 1   INTRODUCTION

     An aerosol is basica'Py a suspension of solid particles and liquid
drops in a gaseous medium.  They are throughout our environment, existing
naturally and as a by-product of man's activities.  Particles with
diameters less than 10 mm are generally classified as aerosols.  This
range of particle size is important in terms of ambient air quality,
health hazards, and industrial pollution control.  In particular, fine
submicron particles, those less than 1 ym, are of most concern with
respect to visible light scattering, atmospheric dispersion, deep penetration
an'! retention in lungs, and decrease in conventional removal mechanism
efficiencies.

     Aerosols can be composed of a multitude of substances depending on
t.ia materials involved in their generation.  This report will be limited
to a discussion of metal aerosols.  The ferrous and nonferrous metallurgical
refining, processing, and manufacturing industries make up one of the
largest as yet uncontrolled sources of anthropogenic air pollution,
second only to power plants.  Because of such large amounts of metallic
compound emissions, it is important to delineate the specific characteristics
of these aerosols.  Continued study of the physical, chemical, and
biological properties of these aerosols will lead to a better understanding
of their health effects as well as capabilities of controlling and
reducing their emission.

     A brief review of some metal particulate emission data will help to
put the significance of the problem into perspective.  Various sources
quote wide ranges of emissions for various metallurgical facilities.
Estimates of uncontrolled particulate emissions include:  8-51 Ib/ton
steel produced1"4, 20-60 Ib/ton Cu4, 50-90 Ib/ton Zn4, 12-75 Ib/ton Pb4,
                   4
and 81-98 Ib/ton Al .  Comparisons with controlled emissions indicate
that 99.5% control is possible for Cu, Fe, Pb, and Zn, giving emissions
                                 4
of approximately 0.1 - 0.5 Ib/ton .  Controlled Al emissions are somewhat
higher, in the range of 1.6 - 19.6 Ib/ton  .  These five metals comprise
more than 95% of total U. S. metal consumption according to a 1976 EPA
report . Small amounts of Mn, Ti02, Ni, Zr, Sn, Sb, and Mo are

                                      1

-------
also used.  Even if emissions  of many  of  the  five major metallurgical
industries are eventually  controlled,  pollution  problems could still
exist because of the effect  of the  particle size distribution on  control
efficiency.

     Although there are  references  in  the literature to metal aerosol
particle  size distributions, they  vary widely,  due  partly  to the  nature
of  the  aerosol.  Fumes are a type  of aerosol  which  result  from  oxidation
reactions,  sublimation,  or condensation,  mechanisms associated  with the
application  of  heat.  Most metallurgical  processes  involve heating and
melting at  some stage, generating  large amounts  of  fumes.   The  mechanisms
that generate  these fumes  generally operate at  the  molecular level,
 resulting in an aerosol  of largely submicron particles.   These  aerosols
 are usually polydisperse,  having  a wide range of diameters.

 1.1  Metal  Aerosols  from the Iron  and  Steel Industry

                                                                   2
      In a review of  the  ferrous foundry industry,  Bates and Scheel
 found that the particle  size distributions varied  from 50% weight  <  1 ym
 to 5% weight <2 ym.   The composition  also varied  considerably, with the
 iron content shifting from as much as  50% to only 5% depending  on the
 impurities present and on the stage of the process.  Other components
 besides iron oxides  (Fe^O^ and Fe-jO^)  in foundry emissions are  Si, Ca,
 Al, Mg, Zn, and Mn.   Electric arc emissions, however, are  mainly composed
 of Fe and Zn oxides  and  volatile matter.   Bates indicates  that  other
 sources such as Holland  et al. and Coulter have found electric  arc
 emissions to be 90-95%  <  0.5 ym.   Another study of electric arc emissions3
 states that particle size measurements by Erickson  indicate 70% weight
 <5 ym, while electron photomicrographs by Allan et al. found 95% by
 number of the particles  < 0.5 ym in diameter.

      Several studies have determined the particle size distributions
 from iron cupolas.  Two publications by Calvert6'7 give results for gray
 iron and ductile iron cupolas:  the mass median diameters  range from
 0.25 to 1.9 ym.  For the Calvert data a more exact breakdown of the
 results is given in Table 1  which summaries data on metal  aerosols from
 the iron and steel industry.  In another study on a gray iron source,

-------
Table 1.  SUMMARY OF PARTICLE SIZE DATA FOR THE IRON AND ?TEEL INDUSTRY
Source
Foundry
Cupola
Gray Iron
Gray Iron
Ductile Iron
Electric Arc
Ferroalloy
Ferroalloy
Author
Bates2
Calvert4
Calvert5
Cooper
Calvert5
Bates2
McCain7
Drehmel8
Percent
by Weight
< 1 ym
50
40 - 60
40 - 80
80
60 - 100
95
20
15
Mass Median
Di ameter
ym
-
0.92 - 1.15
0.54 - 1.9
0.5
0.25 - 0.84
-
4.3
1.7
Deviation
-
1.7 - 2.1
1.5 - 1.8
2.4
1.5 - 2.0
_
-
1.9

-------
Cooper8 f',und a mass median  diameter of 0.5  ym.   For ferroalloy furnaces,
McCain  determined  a mass  median diameter of 4.3 ym and  Drehmel10, one
of 1.9 ym.   It should  be noted from Table 1  that for most  of  the  iron
and steel sources tested submicron sized particles ranged  from 40 to 100
weight percent.  The exception appears to be ferroalloy furnaces  where
the percentage is slightly lower (15-20 weight percent).  For all
sources  the  mass median diameter ranges from 0.25 to 4.3 ym and  the
geometric deviation from 1.5 to 2.4.  If it is assumed that particles  of
3  ym  in  diameter and smaller are respirable, then it may be concluded
that  metal  aerosols from iron and steel operations are concentrated in a
fairly narrow size  range most of which is respirable.

 1.2  Metal  Aerosols from the Nonferrous Metals Industry

      Estimates of aluminum emissions are 40% by weight < 1 ym and 60%
 by weight <  5 ym in diameter.  Measurements by Hanna and Pilat   at a
 horizontal  aluminum reduction cell show 30% by weight  <1 ym, 16% by
 weight  < 0.2  ym, with a mass median diameter (dj) of 5.5 ym  and a
 geometric standard deviation (a ) of 25.  They felt that coagulation and
 diffusion to the exhaust duct wall was a factor in reducing the number
 of particles ,< 0.5 ym and shifting the particle size distribution up.

                        12
      Harris and Drehmel   cite information by Vandegrift showing that,
 on the  average, at least 30% of various metal particulate emissions will
 be less than 10 ym in diameter.   Furthermore, Harris gives the PSD of
 the  emissions  from various nonferrous industrial process units.  For a
 lead sintering machine the d^ was 0.89 ;yra. and the a  was 4.4.  For a
 zinc sintering machine, the d     was nearly the same (0.82 ym) and the
  distribution more  nearly monodispersed  (indicated by a a  of 2.1).   At
  the  same  plant  the zinc roaster  had very fine particulate emissions  as
  indicated by 90% less  than  1 um  and a d    of 0.42 um.  The size distribution
  was  similar with  a  a   of  2.3.   For a copper converter fewer  small
  particles  were  found  and  the  d^ was  the highest of the four  units
  tested  (1.05 ym).   The particle size distribution was also the narrowest
  with a   =  1.7.  A summary of  data on metal aerosols from the nonferrous
  industry is given in  Table 2.   Note that all  these nonferrous  sources

-------
     Table 2.   SUMMARY OF PARTICLE SIZE DATA FOR THE
                    NONFERROUS METALS INDUSTRY
Source
Percent
by Weight
  < 1 yni
                                   Mass Median
                                   Diameter
               Deviation
Al Reduction
   Cell

Zn Sintering
   Machine

PB Sintering
   Machine

Cu Converter

Zn Roaster
     30


     60


     55


     45

     90
5.5


0.82


0.89


1.05

0.42
25


2.1


4.4


1.7

2.3

-------
have at laast 30% of their emissions  in the  submicron range; the percentage
may be as high as 90%.  The d^ range from  0.42  to  5.5ym.  Recalling
that the d^ for iron and steel sources  ranged from 0.25  to 4.3 ym, one
may conclude that the majority  of metal aerosols  generated by  industrial
processes are between 0.25 and  5.5ym in  diameter.

1.3  Metal Aerosols Characterized from Occupational  Studies

     Studies have been conducted to measure  the work environment  of
metal workers and welders to  determine their exposure  potential.
        1 o
Kolderup   studied the fumes  from a ferrosilicon  open  submerged-arc
furnace;   he found basically  a  log-normal  size distribution of agglomerates
ranging from 0.02 to  1 ym, consisting of  spherical  primary particles
ranging from 0.01 to  0.3  ym.  The primarily  silicon  dioxide particles
appeared  to form by homogeneous nucleation between  Si  and  02-   The
agglomerates then grew by diffusion.  Health effects hazards due  to
nitrogen  oxides  (NO  ) and trace elements  such as  Mn  were reported  by
             14                                3
McCord  et al.   , in addition  to the 90-682 mg/m   concentrations generated
with coated welding electrodes. McCord et al. indicated that, although
 a large part of  the  sample may  have been  < 0.5 ym,  the  impinger collection
 system  did not  collect these  small  particles.  When  silica-coated  iron-
 core electric  arc welding electrodes  are  used, particles tend  to  be in
 the size  range  of 0.01 to 1 ym, according to experiments by Pfefferkorn
 and Desler.    By developing  a method  to remove the  iron oxide  from the
 sample,  they observed hollow  Si02 spheres, indicating  that the aerosol
 had an  iron  oxide core with  a Si09  coating.
                                  L.
      In a recent study by Stettler et al.   ,  lung  biopsy specimens from
 two arc-welders were compared with air samples  from their work environment.
 Most of the particulate material  found in  both  samples  was iron,  with
 other components present indicating stainless steel.   Some large  spherical
 stainless steel particles in the range 0.5 -  3  ym  were  found in the lung
 tissues, but most of the particles were clusters 0.1  -  0.2 ym in  overall
 size, made up of individual  particles as  small  as  0.005 ym.  Six  times
 as many irregularly shaped aluminum-containing  particles on the order of
 1  ym were found in Case 2 (the sicker worker) than in Case 1.  Although

-------
difficulties with the filter media precluded sizing the particles from
the air samples, electron micrographs did show that clusters of particles
existed, made up of 0.05 ym primary particles.  The authors point out
that there have been many conflicting reports in the literature as to
the toxicity of iron oxide and other trace materials usually found in
welding atmospheres.  The observations in this study emphasize the need
for further characterization of welding fumes and animal inhalation
studies to quantify their biological impact.

     Concentrations in the breathing zone of operators exposed to two
types of welding equipment, single electrode jet-arc and powder burning
arc, ranged from 20 to 35 mg/m  in five steel plants in eastern Pennsylvania
With local exhaust ventilation, the exposures dropped below the Pennsylvania
Threshold Limit Value (TLV) of 10 mg/m .  No definite connection between
iron oxide fumes and health effects could be drawn from this study.
     A study to assess the hazards to industrial workers exposed to
metallic fumes with a median aerodynamic diameter (MAD) < 1 ym was
                                                18
performed by exposing rats to iron oxide aerosol  .  Rats were exposed
to Fe90, with a MAD =0.3 ym, a  = 1.8, and concentrations up to 700
    3                          "
mg/m .  Although the rats' lung clearance half time was shorter than
man's, the Lung Dynamics Task Group model for man gave a good approximation
of the lung deposition efficiencies in the rat.  The values for pulmonary
deposition are:  35% for 0.3 ym, 25% for 1 ym, and 10% for 5 ym.  Results
also indicated that depositions in organs other than the lung are insignificant
for relatively insoluble particles such as iron oxide.  Casarett and
       19
Epstein  , in another rat inhalation study, found an average deposition
rate of 60% when using a tagged iron oxide aerosol with a count median
diameter (CMD) of 0.068 ym (a  = 1.62).  There was a 3 to 1 ratio between
upper and lower respiratory tract deposition with 10-15% alveolar deposition.

1.4  Characteristics of Generated Metal Aerosols

     A fundamental requirement for any study or project involving the
examination of certain properties of aerosols is a controllable source
of those aerosols.  Laboratory generated aerosols can be used in many

-------
areas of aV pollution research;  e.g.,  biological  studies involving
inhalation and particle deposition;  sampling  instrumentation development,
verification, and calibration;  development  and  evaluation of air pollution
control equipment; and evaluation of industrial  hygiene  ventilation  and
safety equipment.

     Many factors must be  taken into consideration if a  useful  representative
aerosol is to be produced.   The concentration of the aerosol  in the  gas
stream can vary considerably,  depending on  the  method of generation.
The  concentration to  be generated depends on  the needs of the research:
levels less  1 mg/m  would  be useful  for ambient studies; values to
several hundred mg/m  would  apply to industrial  hygiene  and animal
                                                o
inhalation  studies; and concentrations  of 1 g/m  and above  would apply
to process  and  stack  emission analysis  and  control.   Together with the
concentration,  the  gas  flow  rate determines the generation  rate (in
units  of  mass  per  unit  time), which  is  an indication of  a generator's
capability  for  a  certain  application.

      The  particle  size  distribution  (PSD) is  another factor which  is
 radically affected  by generation method.  Different levels  of dispersity
 may apply more  or  less  to different  investigations.   For example,  very
 controlled  studies  (e.g., development of a  respiratory model  or sampling
 instrument  calibration) may  require  a relatively monodisperse aerosol,
 while control  equipment testing and  health  effects research may be more
 realistically accomplished using a polydisperse aerosol, as would  be the
 case in real life.  Shape  and density are two other factors  which describe
 the overall  state  of the  aerosol.

      To generate liquid or solid aerosols  the basic input requirement is
 energy.   In general,  the  more energy per unit of material,  the smaller
 is the particle size attained due to increased surface area.  This  energy
 input can take two basic  forms to generate  either dispersion-redistribution
                                   20
 aerosols or condensation  aerosols.    In the former, liquid aerosols are
 normally dispersed by nebulizers and sprays in various configurations,
 while solid particles can be redistributed  by blowers, grinders, vibrators,
 or resuspension in liquids.   Condensation  aerosols are created by heating
 a material  to the vapor state, by combustion, or by chemical  reaction.

-------
                                                          20-23
     Many good reviews of aerosol generation are available      which
discuss various techniques and mechanisms involved in this important
field. This report will examine these topics as they apply specifically
to metal aerosols.  Although this sometimes involves overlap with techniques
(such as nebulization) applicable to liquids or solids, other techniques
(such as exploding wires and electric arcs) are unique to the generation
of metal aerosols.

     Researchers have been interested in the characteristics of metal
aerosols since the early 1950's.  Advances in generating and analyzing
aeiosols in the past 5 years have increased the desirability of using
submicron diameter metal particles.  The major generation techniques
that will be reviewed are:  exploding wires; electric arc, both low and
high intensity; vaporization; combustion; and liquid resuspension systems
a- -1 associated thermal treatment.  Inherent in many of these methods
will be factors influencing the quality and usefulness of the aerosol.
Such factors include the freshness or staleness of the material, the
moisture content and temperature of the aerosol-gas stream system, and
the presence of combustion products and other gases.

     Condensation processes will be discussed first.  In general these
processes (e.g., exploding wires, electric arcs, combustion, and vaporization)
produce very high concentrations of very small particles, with large
size distributions.  The high concentration promotes rapid agglomeration
by diffusion, which changes the particle size distribution.

2    EXPLODING-WIRE GENERATORS

     The general principle of this method is to conduct as much stored
electrical energy through a thin piece of wire in as short a time as
possible.  Both amount of energy and time of transfer are important
parameters.  If the amount of energy transferred exceeds that which
would cause surface evaporation, the wire expands as a superheated vapor
until it explodes.  The wire thus changes state and becomes a charged
plasma of metal vapor which cools and condenses into a submicron metal
aerosol.  The plasma consists of the metal vapor and ionized gas.
Thermal coagulation and Brownian diffusion occur rapidly, creating

-------
charged chained groups of particles.  A basic series of texts containing
extensive information on the exploding wire phenomenon was compiled by
Chase and Moore   during the period 1958-1968.  More recent literature
reviewing characteristics of the aerosols is presented in a chronological
manner.

     Metal aerosols from exploded wires of 15 different materials were
studied by Karioris and Fish.25  With a current surge from a discharging
4000 joule capacitor, spherical metallic or metal oxide particles with a
CMD of 0.02 ym are formed.  This median diameter was observed to decrease
with an increase in voltage.  Although agglomerate sizing was not done,
coagulation occurred quite rapidly and could be observed directly in a
Tyndall beam.  By diluting in the ratio of 20 to 1 and mixing for an
hour, a stable aerosol of mostly isolated spherical primary particles
                                   7fi
could be attained.  Duerksen et al.   improved upon the previous exploding-
wire generator with construction of a smaller, self-contained unit.
They reported obtaining reproducible spherical particles in the range of
0.01 - 0.1 ym for use in studies of particle properties when exposed to
various humidity, radiation, and gaseous contaminant conditions.  Other
             27
investigators   using the same equipment studied tagged S02 adsorption
rate with aerosols generated from exploded Fe, Al, and Pb wires.  They
reported that chemisorption accounted for a monolayer coating of S02 on
the particles but physical adsorption continued to add up to 75 times as
thick a layer of S02-

                                                        28
     A modification was developed by Tomaides and Whitby   which altered
the energy triggering method.  In their generator, the wire was brought
into contact with a fixed electrode, completing the electrical circuit
and discharging the energy.  They also observed the relationship of
decreasing the CMD and the a  as specific charge energy (energy per unit
length) was increased.  A typical aerosol generated by this process had
a CMD = 0.03 ym and a  = 2.0.  The aggregates formed by these primary
particles, however, experienced an increase in CMD with an increase in
specific charge energy.  It was theorized that the smaller particles
promoted greater coagulation, ultimately resulting in a larger but log-
normal  distributed chained aggregate aerosol.
                                   10

-------
     Phalen29 Investigated the properties of an exploded silver wire
aerosol with particular emphasis on an evaluation of its suitability for
inhalation studies.  He determined the coagulation constant to be 2 x
  -93                                         93
10   cm /sec, with an initial concentration of 10  particles/cm  ,
suggesting a coagulation half-time of approximately 0.5 seconds.  Accordingly,
he aimed a sharp blast of air (2 liters) at the wire cloud, triggered
within a few tenths of a second after the explosion to delay the coagulation.
A large dilution factor then maintained a stable aerosol with a mean
aerodynamic diameter (MAD) of 0.3 \an.  Due to the configuration of the
aggTomerated aerosols with their greatly increased surface area per
vccime, the particles are more chemically reactive at the surface, can
retain more electric charge, and can be more easily solubilized.  These
factors can affect aerosol interaction with biological systems and make
them more hazardous than was previously expected.

     Rather than simply characterizing the aerosol, Sherman   concentrated
on the relationship between the particle size characteristics and the
actual amount of energy transferred to the wire.  With his circuitry and
measuring instrumentation, he was able to determine the energy, in
joules, as a function of time, in microseconds.  Transmission electron
microscope (TEM) analysis of Ag, Cd, and Zn aerosol collected on mica
coated plates (positioned 8 cm from the wire) indicated that the mean
particle size decreased as energy transfer increased.  That is, the
faster the metal expansion, or explosion, the greater the supersaturation
and rate of nucleation. He observed that Cd and Zn had a narrower size
distribution than  the Ag aerosol.  This may be explained by the fact
that Ag, having a  larger surface tension, would tend to have a slower
nucleation rate yielding a wider size distribution of nuclei.  It is
interesting to note that Sherman saw no change in PSD over several
minutes—the electron photomicrographs showed that almost no agglomeration
took place.  His major observations (when no chemical reactions occurred
and pure metal aerosols formed) were:  particles were spherical in shape
and ranged from 0.005 to 0.1 ^m in diameter; the PSD was roughly log-
normal; and the mean particle size was inversely proportional to the
energy transferred per unit length of wire.

                                        n

-------
     Just as the energy  transfer  that  Sherman  studied is one of the most
critical parameters  in determining  aerosol  characteristics prior to the
explosion, the phenomenon  of  agglomeration  or  coagulation, seen by most
experimenters, is  one of the  most critical  parameters after  the explosion.
       31
Wegrzyn   studied  the effect  the  bipolar charge has on  the stability  of
the generated aerosol.   Using a gold wire,  exploded in  an  argon-filled
chamber, he  observed an  aerosol whose  primary  particle  mean  diameter
ranged  from  0.01 to  0.1  ym with a typical effective diameter of 0.032 ym.
The percentage of  charged  particles was slightly less than 50%.  The
charge  concentration of  his aerosol (approximately 100,0007cm ) was well
within  the  capabilities  of a  2 millicurie Krypton-85 neutralizer at a
 flow  rate  of 10  1/min.   Wegrzyn then compared  the concentration in
 particles/cm  over a 30  minute period  after the explosion with and
 without exposure to the  neutralizer to observe the coagulation coefficient
 (K) with respect to time.   After  the aerosol was neutralized to a standard
 Boltzman distribution of charge,  K was fairly  stable within  a range of
           -10   3
 21-29 x 10    cm /sec; however, without neutralization, values of K were
 approximately a factor of 2  higher during the  first 600 seconds, before
 dropping to the lower values.  He also noticed an increase in the total
 mass concentration  with respect to time which  may indicate the increasing
 vapor adsorptive capability  of the coagulating aerosol.  His results
 reinforce previous  findings  that exploding wire aerosols are reproducible
 (submicron, normal, or  log-normal) in primary particle size distribution
 and highly  charged.

      Radioactive  tagged iron oxide aerosols from an exploding wire were
                         32  33
 analyzed by Kops  et al.   '    with  an electron microscope and a Stober
 centrifugal aerosol spectrometer (which deposits the aerosol along a
 strip depending on  the  aerodynamic diameter)  to study fluid drag on the
 chained aggregates.  The  electron  microscope  examination showed that the
 primary particle  size distribution was approximately log-normal.  From
 the centrifuge data of  iron  oxide  aggregates  with six different primary
 PSD's, they consistently  observed  two distinct log-normal distributed
 diameter ranges which depended on  the concentration (n) in the chamber.
                3    3
 For n  > 5  x 10 /cm , the  aerodynamic diameter (dfl) was proportional to
 n1/3;  for n< 5 x 103/cm3, da was  proportional to n1/6.  Typical mean d
 values of the two regions  were 0.259 and 0.925 ym.  This is as expected,

                                     12

-------
since a higher concentration would tend to promote more coagulation  and
a wider subsequent diameter spread; whereas, a lower concentration would
have a lower coagulation rate and, hence, a more monodisperse aerosol.
Their results are of importance in describing the coagulation of solid
particles.  It has generally been assumed that smaller solid as well as
liquid aerosols unite to form one large new particle, implying that  the
Stokes diameter of a chain aggregate equals its volume equivalent
diameter. Kops et al., however, from their calculations of d . have
                                                            a
shown that the Stokes diameters of chain aggregates are less than their
volume equivalent diameters.

     Exploding wire aerosol generators have been utilized and studied
extensively for the past 30 years.  Many of the phenomena are fairly
well understood and, with the various refinements that have been incorporated
in their  use, the aerosol is quite useful and dependable in the proper
application.  The inherent batch operation nature of exploding wires,
however,  limits its usefulness when larger, more continuous quantities
of aerosol are desired or required. Moreover, as shown in Table 3, the
size distribution of particles generated by exploding wire techniques is
very small (0.005 - 1 ym) and does not simulate industrial metal aerosols
which range from 0.25 to 5.5 pm.

3    ELECTRIC ARCS

     One  metal aerosol generation method that is not limited to small,
batch type conditions is the electric arc.  The technique of striking an
arc between two electrodes has been recognized since the late 19th
century.  Continued  development led to the discovery of the high intensity
arc mode  by Beck in Germany in 1910.  Early aerosol scientists such  as
            34
Whytlaw-Gray   were making use of electric arcs to study the properties
of metal  fumes in the early 1920's.  Ultramicroscopy allowed those
researchers to observe particles in the 0.01 - 0.2 um range.  With the
advent of electron microscopy and other advances in particle sizing  and
counting  methodology, the study of electric arcs continued with emphasis
on the characterization and use of the aerosol being generated.  Research
with both low and high intensity arcs has contributed to the understanding
of metal  aerosols.
                                     13

-------
 Table 3.  SUMMARY OF PARTICLE SIZE DATA FROM EXPLODING-WIRE GENERATORS
Author
                  Data Type
                              Result, ym
Karioris
Duerksen
Tomaides
      2Q
Phalen"
Sherman
Wegrzyn
     32
     ^
 26
 28
30
31
Count mean diameter
Range
Count mean diameter
Mass average diameter
Range
Effective diameter
Mean diameter
0.02
0.01 - 0.1
0.03
0.3
0.005 - 0.1
0.032
0.26, 0.93
                                 14

-------
3.1  Low Intensity Arcs

     A low intensity arc normally consists of one or two  consumable
electrodes creating a gap across which an electric current can  be  passed
in a continuous spark.  A great deal of the electrical energy is transformed
into thermal  energy, which melts and partially vaporizes  any material
within the arc with a melting or boiling temperature less than  the arc
temperature.  Typical arc-spot temperatures (actual spot on metal that
arc touches)  are approximately 3000°C.  The exact configuration of the
apparatus will determine the dispersion of the melted and vaporized
metal.  As the material leaves the arc zone, it rapidly cools,  causing
melted drops to solidify and vaporized material to condense and solidify
into an aerosol.  The gas in the arc zone between the anode and the
cathode becomes ionized and exists as a plasma.  Since the majority of
the voltage drop occurs in this region, most of the energy transfer is
in the form of radiant heat to the plasma rather than conductive and
convective heat to  the electrode which is to be melted. This condition
of energy distribution is optimized with the high intensity arc which is
discussed later.  The aerosol generated by the low intensity arc is
influenced by many  input parameters such as voltage and current applied
to the arc, rate of electrode consumption, and temperature in the  arc
zone.

     Since the consumable electrode is the initial source of aerosolized
                                                              35
metal, its physical characteristics must be understood.   Amson   developed
a model for predicting the voltage along the "stick-out"  (the exposed
electrode between the current pick-up point and the arc interface) in
terms of arc current, wire feed rate, and exposed length.  Carslaw and
Jaeger   also present a well-developed model of temperature distribution
within the electrode.

     Recognizing the need for the generation capability of aerosols<1  ym
from an industrial  hygiene standpoint, a group of Japanese investigators
generated Fe, Pb, and W aerosols with an electric arc.  Rather  than
using a continuous DC current, they discharged 9000 volts AC intermittently
(300 cycles/minute) across the electrodes.  Sampling with an electrostatic
precipitator sampler and analyzing with an electron microscope  indicated

                                     15

-------
spherical Fe and Pb oxide  particles with  little agglomeration.  The
count mode diameter of  the iron  oxide was 0.02 ym; for the tungsten, it
was 0.04 ym, with a general  range  of 0.015 -  0.15 ym.  They concluded
that this was a viable  method  of generating submicron  particles, but did
not attempt to explain  the formation mechanisms.
                  OQ
     Harvey et al.    generated many metal aerosols  with  a DC  arc to
investigate the structure  and  forms of  the primary  particles  of the
smokes.  Electron microscope analysis confirmed earlier investigations
 (such  as by Whytlaw-Gray et al.   ) that all the particles formed charged
 chain-like  aggregates.   Additionally,  the extent of the  chain formation
 for Au and  Ag  smoke appeared to  be strongly dependent on the pressure of
 the gaseous  atmosphere. The Au  and Ag  aerosols (from heated wire filaments)
 were formed of pure metal; the other metals investigated formed oxides,
 except that Pb and  Bi formed metal particles coated with an oxide.  From
 electron and x-ray  diffraction analysis,  some typical  shapes, and primary
 and aggregate dimensions are:   ZnO--hexagonal, 4$ (lft =  10" ) primary
 particles,  and 4000$ long  rod-shaped  aggregates;  Fe203~ cubic, 8$
 primary particles,  and 300$ aggregates; and Al£0,—cubic, 16$ primary
 particles,  and 300-500$ agglomerates.

      Although researchers  had made many estimates of aerosol  shape and
               39
 size, Reichelt   was among the first  to critically examine the operational
 parameters and their influence on the  aerosol particle shape and size
 distribution.  Experimenting with Fe,  Al, and Cu electrodes in air, CO^,
 argon,  and ^» he could detect no lower size limit with  the capabilities
 of the  electron microscope (size range from 100$ to 20 ym).  These
 particles are created  as  some of the  metal evaporates due to the high
 temperature at the spot where the electric arc contacts  the electrode.
 Even if voltage and current are maintained constant, the arc cannot be
 considered a candidate for use as a monodisperse aerosol generator due
 to the  nonuniform nature of the arc,  which can move about quickly,
 changing its length and spot location.   In terms of total aerosol  production,
 current has a great influence:  the weight of aerosol  is proportional  to
 the current cubed.   Thinner electrodes and greater separation optimize
 and increase production.  Analyzing particles with a Goetz aerosol

                                     16

-------
centrifuge, Reichelt found that the particles generated were not spherical
and formed chains.  He also noted that the current had little effect on
the Stokes diameter (diameter of a sphere with the same density and
settling velocity as the particle in question) particle size of the
chain aggregates.  Increasing the flow rate did decrease the agglomerate
size, basically by diluting and dispersing the particles, reducing the
coagulation rate.  With electron microscope analysis it was found that
current had a large effect on the size of the primary particles—an
increasing current increased the size. This tends to explain the fact
that the influence of current on Stokes diameter is not so pronounced,
since larger particles will coagulate less. Other aerosol characteristics
(e.g., the number and size of polyhedra and crystals) varied with flow
rate, direction of flow (horizontal or vertical), and chemical composition.
In conclusion, Reichelt felt that the arc was not an ideal aerosol
generator, due to the rather uncontrolled and complex processes involved
and his results of varying PSD's with consecutive samples.

     Beginning with a configuration similar to Reichelt1s with two
                                                       40
electrodes inserted into an exhaust duct, Linsky et al.   investigated
the generation of iron and zinc oxide.  The initial generator consisted
of one consumable wire electrode and one nonconsumable tungsten electrode.
Results indicated that 75% of the agglomerated iron oxide particles were
 <0.1 urn; primary particles had a mean diameter of approximately 0.02 ym.
The agglomerate  size data probably was altered since the measurement
technique was resuspension in liquid for Coulter Counter sizing.  Linsky
et al. later examined  a commercially available wire spray metal!izer,
used to coat industrial parts with a fine metal layer.   The equipment
consisted of two consumable wire electrodes mounted on a gun; a compressed
air nozzle centered between them dispersed the melted and vaporized
material.  This configuration was much more controllable than the earlier
one or Reichelt's equipment, and reproducible PSD's and concentrations
were obtained.  It was estimated that agglomeration occurred within the
first 0.1 second after generation, forming highly charged chain aggregates
whose mean diameter was 0.41 ym; the mean diameter of the primary particles
was  <0.01 ym.  The maximum aerosol generation rate of the Linsky metallizer
was approximately 8 g/min which corresponded to a typical conversion
efficiency of wire metal into usable aerosol of 8-15%.
                                      17

-------
     Zimm^man41achieved  similar results using  identical equipment, with
an approximate conversion efficiency of 4% and  a highly charged, agglomerated
aerosol.  Approximately 80% of this aerosol had an aerodynamic diameter  <
1 ym, based on a  cascade  impactor.   More aerosol characterization work
needs to be done  on  this  system to  determine if a specified set  of  input
parameters  (e.g., voltage, wire feed rate, compressed air pressure,
exhaust air flow  rate,  and wire composition) can generate a reproducible,
predictable aerosol.

 3.2   High  Intensity  Arcs

      A high intensity arc is similar in configuration to the low intensity
 arc.  The  difference lies in the current density and efficient  use of
 the electrical  energy.   The high intensity arc is established at a
 sudden transition point when the current density exceeds a critical
 value for an ordinary arc.  In the arc crater, the temperature  can
 increase by more than a factor of 2 to 7000-8000°C.  As previously
 mentioned, in the low intensity arc the largest voltage drop occurs in
 the plasma between the electrodes,  while the remaining drop is  approximately
 10% across each  electrode.  As the arc current increases, the current
 density remains  approximately constant as  the arc crater (assuming
 electrode diameters larger than wires) increases on the surface of the
 anode until it fills the  entire anode face area. Additional energy is
 dissipated by radiation  and conduction of  heat until the increased
 energy can't be  removed  fast enough.  Then the anode temperature begins
 to  rise until the boiling  point is reached.  Beyond this point  large
 amounts of vapor are generated by the anode material.  A superheated
 region develops  near the anode in this tail flame which can reach values
 to  twice the boiling point of the anode material  (e.g., 4827°C  for
 carbon; 2750°C for  iron),  dissociating  it  into elemental molecules which
 then  recondense  as  they  travel out of the  hot zone.

       The high intensity  arc was proposed by Sheer and Korman42  as a
 means of energy  transfer in the processing of materials, such as refining
 of  metal ores.   The nonconducting portion  of the  electrodes can be as
 high  as 80-85%,  with the balance being  carbon.  They obtained silica

                                       18

-------
fumes with an average particle size of 0.03 - 0.04 ym with an  80%  silica/
20% carbon electrode.  Harris et al.43 later proposed the use  of the  arc
for processing rhodonite for extraction of manganese from siliceous
ores. No size data was given in their study.

     Holmgren et al.  applied the high intensity arc principle specifically
to submicron particle formation for industrial applications.   By consuming
the solid anode, a highly ionized plasma anode tail flame exists,  so
that there are many available condensation nuclei upon which submicron
particles can form.  The rapid temperature drop does not allow much
particle growth leaving primary particles in the 10-1000$ size range.
Increasing the dilution-cooling quench rate decreases the size.  Increased
quenching also decreases the initial coagulation rate by dispersion of
the particles and thus limits the PSD.  Agglomeration, however, cannot
be avoided, and chain aggregates will form.  According to the  authors,
the increased chemical reactivity, sintering rate, and surface energy of
submicron particles make this a promising process.  For the same reasons
it also could be very useful in aerosol generation applications.

     Although the anode temperature of a high intensity arc can reach
8000°C, the area around the cathode can achieve temperatures in the
                      AR dfi                                          d
-------
     Researchers in Italy  (Boffa  and  Pfender47) have made a similar
application of the high  intensity arc principle with the added variation
that the quench gas transpires  through the anode  as the anode is ablated,
or consumed.  This transpiration  not  only provides dilution and cooling
air but appears to limit coagulation  as well.  Their results show that
they can control mean  particle  size between 458  and  1  ym with o    <
                                                               g
1.2.  A a   this low is generally  considered to be a  measure of practical
monodispersity.  The major controlling parameters are  the  arc current
and the cooling gas flow rate.  The mean particle size increases  with
increasing temperature,  due to  increased current, because the anode
ablation  rate increases  causing a greater density of particles  in the
area  near the anode.   This increase in number of particles influences
the coagulation  and  growth rate.  The mean particle size decreases,
 however,  with an  increase  in flow rate due to enhanced cooling and
 dispersing.

                  48
      Boffa et al.    later altered their generation system to make it
 easier to use.  The high intensity arc is maintained between two nonconsumable
 electrodes with argon gas feeding into the cathode area.   The plasma jet
 and the argon transpire through a porous consumable matrix.  The plasma
 ablates the matrix material while the gas quenches and dilutes the
 vapor.  Similar PSD results were obtained with this method.  No generation
 rate information was  given for either variation of this method.

                                               49
      Another group of Italians, Tarroni et al.   , developed an apparatus
 very similar  in principle  to the work of Korman et al.    Tarroni's
 group utilized a Collison  atomizer (with dialyzed Fe^O., suspended in
 argon) and a dust feeder  (with iron  particles) to feed the material  into
 the  plasma arc near the cathode.   CMD's of 0.005 and 0.004 ym, respectively,
 each with o  = 1.3, were  obtained.

                                  50
      A review by Fannick  and Corn    indicates the extent of industrial
 hygiene  hazards of plasma  torches  and  arcs.   Concentrations of various
 fumes generated at plasma  torch  facilities were presented.  Many exposures
 of Fe203  exceed the Threshhold Limit Value (TLV) of 10 mg/m3, ranging
 from 11.9 to  35.1 mg/m  .   From another  source, they reported total fumes
                                                            o
 (consisting  of 10% chromium) ranging from  24.5 to 31.3 mg/m  without

                                        20

-------
exhaust ventilation and 0.4-5.7 mg/m  with ventilation.  Although  this
data is presented from an industrial hygiene standpoint, it  indicates
the generation capability of high intensity plasma arcs in industry.

4    HEATING TO EVAPORATION

     Rather than imparting electric energy to the metallic material as
quickly as possible in a violent action, such as the electric arc  and
exploding wire, other researchers have studied the effect of applying a
steady source of heat:  dirdctly, as with a lower level of electric
current; or indirectly, as with a furnace chamber.  This also results in
the vaporization of a steady stream of metal vapor which condenses into
a submicron aerosol. Some controlling parameters include temperature of
the material, composition and dimensions of the material, and dilution-
carrier gas flow rate.  In addition to the development of practical test
aerosol generators, the purpose of many of the following studies was to
further investigate the underlying principles involved in the evaporation/
condensation mechanisms.

              51
     Turkevich  stated his purpose as not only examining the nature of
air contaminant fumes, but also, from an acadmeic standpoint, determining
how unordered gaseous atoms condense to ordered ultrafine particles.  In
the region  in which molecules aggregate in sufficient numbers to promote
condensation, theoretical and experimental evidence seems to indicate
that a limiting lower particle size exists, based on the fact that
surface tension (varying directly with the square of the radius) tends
to disrupt  the particle while lattice energy (varying directly with the
cube of the radius) tends to hold the particle together.  Thus, the
point of formation of a particle is complex, involving many  parameters.
For his purpose, Turkevich felt that the particles produced  by an  exploding
wire were too irregularly shaped. He attributed this to the  nature of
the particle formation, which he considered to be basically  disintegration.
A process which agglomerates or condenses is considered more uniform for
the study purposes.  Particles from two aggregation processes, DC  electric
arc and electrically heated wire, were examined.  The smoke  from the
electric arc suggested that the shape of the submicron particle was
often characteristic of the chemical composition of the parent material,
                                      21

-------
especially as the particle  size  decreased.   Turkevich attempted to study
the stages in the evaporation/condensation  process  by passing current
through a tungsten wire  enclosed in a vacuum chamber bell  jar, and
collecting particles  or  promoting condensation on collection surfaces
(collodion grids, later  used directly in electron microscope analysis)
positioned at different  distances from the wire.   With  a high  vacuum
(10  mm Hg)  in  the  chamber, the  vapor appeared to migrate to  each surface
before condensing on  it, giving  a similar uniform deposit on  each,
regardless of distance.   As the  pressure in the chamber increased,
however,  he  was able  to  detect three distinct areas where different
processes predominated:  evaporation, nucleation,  and coagulation.
Evaluations  of  particles were based on the average size and the PSD.
From his  observations he concluded that the final diameter of the particle
 is inversely proportional to the rate of nucleation, a phenomenon observed
 by others.
                                   ro_CT
      Spumy and various  co-workers  "  investigated the generation of
 metals and metal compounds by sublimation and vaporization.  Spurny
 prefers condensation over spraying as a means of aerosol generation,
 stating that the former results in a more monodisperse, controllable,
                         CO
 and reproducible aerosol   .  Two methods which Spurny uses in most of
 his research are: 1) heating metals in a furnace (usually preferred for
 low melting point metals); and 2) passing current through a wire.  The
 major difference between these methods is the time necessary to achieve
 a stable aerosol.  Furnaces can take hours, while heated wires achieve
 steady-state conditions in a matter of minutes.  In general,  condensation
 methods  can produce  aerosols within a size range from 0.001 to 1
 Dilution to a  concentration
 coagulation of the aerosol.
                                 4
Dilution to a concentration of 10  particles/cm  is necessary to prevent
     Heicklen and Luria   presented a review of experimental  results on
kinetics of homogenous particle nucleation and growth.   They  acknowledge
that metal oxide data in this area is limited.  The authors indicate
three means by which heterogeneous nucleation can form a homogeneous
particle: 1) condensation on an ion; 2)  condensation on a neutral  nucleating
particle; and 3) coprecipitation of two  or more species.  Homogeneous
nucleation occurs when a critical  supersaturation is reached  or a  gas

                                      22

-------
phase reaction occurs.  Metal ions react with 02 to give metal oxides.
Some rate coefficient data is given for these reactions.

4.1  Metal Oxides

     In studies conducted by Maiwald   and Polydorova   , a 0.5 mm
tungsten wire was vaporized by heating the metal to 850°C in an air
flow.  A problem inherent to this approach is the continual decrease in
wire diameter which increases the current per unit length unless the
voltage is regulated. Presence of an oxide layer appears to increase
tf    The oxygen content of the carrier gas could be decreased to limit
the formation of oxide but then the gas stream's usefulness for inhalation
tests would be affected unless 02 additions were made.  Results from EM
analysis indicate that particles range from 0.03 to 0.5 ym and are
coagulated.  An increase in temperature caused an increase in particle
size due to the increased coagulation.

     A similar study was also conducted by Polydorova   with platinum
wire. At high wire surface temperatures (1330-1510°C), platinum oxide
forms on the surface.  Since the vapor pressure of the oxide is less
than that of the pure metal, platinum oxide aerosols are generated.  The
higher vapor pressure at higher surface temperatures generates larger
particles in higher concentrations.  Keeping temperature constant, an
increase in air flow dilutes the vapor and reduces the coagulation rate.
                                r     e             T
Aerosols in concentrations of 10  - 10  particles/cm  with size ranging
from 0.018 to 0.048 ym were generated in low air flows  (3-6 1/min).

     Fumes of Fe, Zn, and Cd oxides were generated by heating small
                                                  rn
amounts of the metals until vaporization occurred.    These simulated
welding and brazing fumes were used to compare the sampling efficiencies
of two industrial hygiene filter samplers.  All samples had data showing
that most particles were  < 1 ym.  The authors cite work by Jarnuszkiewicz
indicating that the major portion of welding fume is  <0.5 ym.
                                      23

-------
4.2  Metal Halides

     Metal aerosols,  specifically Ag,  have been widely  used in the study
of ice and rain  nuceleation.   Requirements for an aerosol for this
application  include  being  continuous,  steady-state,  and free of combustion
products  and oversized  particles.  Barchet and McKenzie63 generated
silver iodide aerosols  by  electroplating nichrome wires with Ag,  exposing
them to  I vapor  to  form a  Agl coating, and then heating the wire  by
passing  current  through it.   Particle  production is controlled by sublimation
from the wire and recondensation in the quench air.   A broad range of
particle sizes was  obtained,  with 7%  <0.1 ym and a geometric mean
diameter of  0.01 ym.  They estimated that approximately 50% of the small
particles were lost by coagulation and diffusion by the time the sample
 reached  the  condensation nuclei counter.

      Another Agl aerosol generation approach was taken by Poc and
        64
 Rolleau   to attempt to eliminate all  impurities from possibly influencing
 the ice nucleating mechanism.  A two-step procedure was used:  first,
 passing pure dry N2 over a crucible of Ag heated to 950°C to generate a
 pure Ag aerosol; and second, exposing  this aerosol to a steady flow of I
 vapor. Electron microscopy indicated an Ag aerosol with a concentration
      /•             
-------
particles formed in the first furnace.  Spumy also introduced various
materials to the double furnace as fine powders in a gas stream of dry
air or helium.    An Fe2CL aerosol with a size range from 0.002 to 0.02
urn was produced with a vaporizing temperature of 450-800°C, well below
the melting point (1535°C).  The range of generation rate capability of
this method is not given but Spurny stated that these aerosols could be
useful in experimental aerosol physics and chemistry studies, filtration
studies, and biological research.  As a variation of the heated wire
method, Spurny et al.  *  heated a hollow tube of platinum by passing an
electric current through it; dry air flowing through the center of the
tube carried away the evaporated metal.  They again observed that the
particle size increased with increasing temperature, but decreased with
increasing gas flow rates.

4.3  Metallic Aerosols in Argon

                                                              65
     A heat-pulse cloud chamber was used by Buckle and Pointon  to
study the condensation of metallic aerosols in argon.  A highly supersaturated
vapor is produced by passing a current through a coil to flash-heat the
metal sample above background temperature.  The vapor rapidly cools and
condenses to form a suspension of droplets or solid particles.  Zn and
Pb aerosols were produced in this way, but difficulties were encountered
with  Ca and Cd.  The authors feel that the formation of micron-size
metallic aerosols is attributed to nucleation and growth under conditions
of large temperature and concentration gradients near the supersaturator.
In continued studies of Cd aerosol formation, Buckle and Pointon  '  observed
that the particles were shaped like prisms and spheroids with a variety
of surface features; e.g., dished areas and concentric rings and layers.
This roughened platelike surface effect was thought to be caused by room
temperature condensation of additional vapor onto the aerosol particles.
They noted that the PSD was constant and log-normal with a geometric
mean of approximately 2 ym, regardless of particle configuration.

5    COMBUSTION

     Melting and vaporizing metallic materials can be achieved by means
other than heating; e.g., combustion, feeding the material into a flame,

                                      25

-------
or combining it with  combustible materials  and  igniting the mixture.
Methods of generating metal  aerosols by combustion are more varied than
the previously discussed  methods since there are many ways of introducing
the metal material  into the  flame.   These methods would still be classified
as condensation,  however, since the original phase,  whether liquid or
solids, if vaporized  in the  flame,  then followed by condensation.

      The  problems of  removing submicron-size iron fumes  from  open-hearth
                                         CQ
furnaces  were  studied by  Billings et al.   in the mid-1950's  using
combustion  to  generate Fe^O., for laboratory tests.   Actually, two methods
were  used:  1)  flake iron  powder was injected into the air side of an
air/oxygen/acetylene  flame where it was oxidized exothermally; and 2)
 iron  carbonyl  was burned  in an oxygen/acetylene flame.   The second
method provided a somewhat more controlled generation.   Concentrations
                                                     o
 from these procedures ranged from 0.04 to 0.50 gr/ft .   In field tests,
                                                              3
 they found furnace emission concentrations of 0.02-0.34 gr/ft .  A
 particle size analysis of the furnace emissions indicated a mass median
 diameter of 0.65 - 0.82 ym, while the CMD was 0.047 - 0.057 ym.  Thus
 even in earlier research with metal aerosols, the necessity for simulation
 of actual emissions was recognized.

      Almost 20 years later another group of researchers, Brain et
 al.   , modified  the  same equipment to continue the study of Fe203
 aerosol generation from  iron pentacarbonyl.  They bubbled Np through the
 Fe(CO)r and decomposed the vapor in a 500°C furnace cylinder.  A concentration
            3
 of 340 mg/m  was achieved, composed of an agglomerated polydisperse
 aerosol with a CMD of 0.4 ym (a  = 2) and a mass median diameter of
 1.6  ym. The primary  spherical particles which make up the log-normal
 distributed agglomerates range  in size from 0.01 - 0.03 ym.  The authors
 estimated that the agglomeration occurs rapidly  (with a half-time of 0.1
 second)  in the furnace.  Keeping the furnace temperature above 500°C
 greatly  reduces  the  toxic fumes and CO, making the aerosol suitable for
 animal studies.

      A generator to  produce  uranium fume, U.OP, was devised by Glauberman
            70
 and  Breslin   in the mid-19501s.  A bed of  uranium chips supported on a
 slowly rotating  wire mesh turntable was burned  in an argon/oxygen gas

                                        26

-------
stream.  The authors demonstrated that the median particle size could be
pre-selected in a range of 0.05 - 0.2 ym (a  of 2.5-3.1) by establishing
the appropriate argon to oxygen ratio and total gas flow rate.  Their
particular chamber design allowed sustained generation for 6 minutes.

     Formenti et al.  , working mainly with Ti02, prepared metal oxide
aerosols by carrying the vapor of metal chlorides into an oxygen/hydrogen
flame.  They observed both spherical and polyhedral (crystal) nonporous
particles in the size range of 0.01 - 0.2 vm.  The major parameters
influencing size and shape include gas flow rate, chloride concentration,
                                                                   72
flame temperature, and residence time in the flame.  Juillet et al.
continued the research with many other metal chloride solutions.  Both
groups noted that the aerosol exhibited photocatalytic properties in the
presence of ultraviolet light, properties not found in large or porous
particles.

     Larger quantities of material can be generated by spraying a dilute
metal or metal salt solution in an atomizer burner.  Dharmarajan and
    73
West   used this method to generate metal oxide aerosols in the size
range of 0.1 - 10 ym.  Additional work is needed to characterize the
particle size, to determine the controllability of the particle size
distribution, and to eliminate the soot that accompanies the metal
oxides.

     A similar approach was taken by Carroz et al.   to demonstrate
continuous  high mass flow rate generation of fresh, fine, inorganic
particulate aerosols, to simulate emissions from coal power plants,
electric arc and basic oxygen furnaces, and Zn smelters.  The fume
generation apparatus consists of a solvent burner with a spray nozzle,
pressurized tank, and ductwork.  Inorganic metallic oxides are generated
from the burning flammable solutions of metal nitrates and other salts.
The hot gases and excess air reduce agglomeration. They have obtained
                                     3                         9
particle loadings as high as 16.8 g/m  and concentrations of 10
particles/cm .  Analyses with an electric mobility aerosol analyzer,
optical counter, condensation nuclei counter, thermal precipitator,  and
electron microscope have indicated that the aerosol consists of hollow
spheres (approximately 20 ym) and chain aggregates with mass median
                                       27

-------
diameters of approximately  1  ym.   Methods  were  developed to vary S02
concentration and  particle  resistivity to  better  simulate power plant
emissions over a wide  range of operating conditions.   The method is very
flexible and can be  used  to generate aerosols of  many different oxides
and chlorides.  Data from all  methods is summarized in Table  4.

6     DISPERSION

      Various approaches to  aerosol generation can be classified as
dispersion  processes.   The  basic difference between dispersion and
condensation categories is  that the former involves mainly physical
processes;  the  other is related more to chemical  processes and phase
changes.   Dispersion encompasses the formation of two types of aerosols,
solid particles and liquid  droplets.

      Dispersion of solid particles, by grinding or resuspension of
 pulverized material, results in a highly polydisperse aerosol, with
 doubtful reproducibility and control of size, shape, and concentration.
 Brain et al.    cite mid-19501s references of the use of a Wright  Dust
 Feeder and a commercial blender to resuspend iron oxide and iron  powder.
 Problems stem from the fact that the closely packed, sometimes charged
 particles tend to remain agglomerated and do not represent actual emissions.

 6.1  Liquid Droplets--General

      Generation of  liquid  droplets can  consist of four basic types:  1)
 nebulization; 2)  spraying;  3) rotation (spinning disks); and 4)  vibration.75
 These methods can be  used  with many  kinds of solutions including  those
 with a dilute concentration of solids suspended  in the liquid.  The
 droplets also can be  treated  after formation such as by heating to
 evaporate  liquid, leaving  a solid residue, or  by thermal decomposition
 to alter the chemical  composition of  the material.   By combining  the
 many alternatives of  dispersion generation,  numerous aerosol generation
 techniques have been  devised.  Examples of some  of these typical  applications
 which generate metal  aerosols  are presented  in this  review.
                                       28

-------
            Table 4.   SUMMARY OF PARTICLE SIZE DATA FOR GENERATORS
                                   USING COMBUSTION
Author
Data Type
Result, ym
Billings68
Brain69
          70
Glauberman
Formenti
Dharmarajan
Carroz
      74
Mass median diameter
Count   "
Mass median diameter
Count   "
Median particle size
Range
Range
Mass median diameter
0.65 - 0.82
0.047 - 0.057
1.6
0.4
0.05 - 0.2
0.01 - 0.2
0.1 - 10
1.3
                                      29

-------
                            75-77
     A number of researchers      have tested spinning disk aerosol
generators using an iron oxide solution.  A fairly monodisperse aerosol
is attainable but only for diameters  larger than  1 ym.  Albert et al.75
were able to generate iron oxide particles in the range of  1-8 ym.
Smaller, less monodisperse aerosols can be made by utilizing the satellite
drops rather than the larger drops  (this wastes 90%  of the  mass).  The
disadvantages are that the particles  need heat to stabilize their  structure
and that the particles cannot be used directly for inhalation studies.
Because the air stream contains solvent vapors, the  aerosol must be
collected and then redispersed, making this an awkward research tool.

      Spertell and Lippman76 also had  difficulties, producing only  200
            o
particles/cm  and being  unable to maintain a stable  test  aerosol for
long  intervals.  Even using an aerosol concentrator,  they could not
generate  a  usable aerosol  less than 3 ym.

      Caplan et  al.,   like Albert et  al., were able  to generate a  fairly
monodisperse  aerosol, but  only in the size range  of  2-6  ym.

      6.2  Using Liquid  Droplets—Nebulizers

          A Lauterbach  nebulizer driven by compressed air was used by
                     19
 Casarett  and  Epstein    to  aerosolize  a suspension of FeO.,.  ps^ data
 determined with electron microscope analysis  indicated  a  CMD =  0.068 ym
 and o  = 1.62.   The authors  concluded  that Fe20~  was  an adequate material
 for use in inhalation studies  although continued  work is  needed to fully
 characterize the aerosol, especially in the area  of solubility.

                       78
      Waite and  Ramsden   describe an ultrasonic nebulizer which generates
 radioactive, stable, insoluble particles of Fe^O.,.   A solution  prepared
 from ferric chloride and ferric nitrate results in  hollow spheres of
 Fe203 when the  solution is nebulized and dried.   These  hollow spheres
 have a CMD of 0.03 - 0.5 ym  when the initial  solution droplets  are 3 ym.
 The authors note that an aerodynamic median diameter (AMD) rather than
 CMD should be used to describe the diameter of the  iron oxide aerosol
 since the spheres are hollow rather than solid.   To make  this conversion,
 the density of  the aerosol particles is needed;  it  will be less than the
                                      30

-------
bulk density of iron oxide or of a solid iron oxide particle.  Their
observations show a significant drop in density between 0.55 and 0.3 ym.
This process is not only interesting, but also useful in situations
where the amount of material needed is limited.

                             79
     The work of Raabe et al.   appears more promising but still only
allows a small amount of aerosol to be generated.  Solutions of metals
in the form of oxalates, citrates, and tartrates are nebulized into
droplets and then thermally decomposed in a furnace at 1150°C.  A spherical
oxide aerosol results with no organic solvent contamination.  A zirconium
oxide aerosol was generated and found to have a size of approximately
3 ym.  The authors state that the more monodisperse the droplet distribution
system is, the more monodisperse will be the final particles.

                                                          80
     This project was continued by Kotrappa and Wilkinson,   using the
Lovelace nebulizer, which the authors classify as a high specific output
nebulizer.  They generated droplets from a colloidal iron hydroxide
solution which was nebulized and dried to form iron oxide particles in
the range of 0.3 to 3.5 ym, with a particle material density of 2.56
g/cm  (density was independent of particle size).

     Aqueous solutions of metallic carbonates, acetates, and oxalates
were generated with Lovelace and Collison nebulizers by Horstman et
   81                           82
al.   and Friedman and Horstman.    Cd, Ni, and Mn solutions were used.
The droplets were dried in a furnace with temperature of 600-1000°C to
decompose them to an aerosol with a mass median diameter of 0.15 -0.6 ym
and a  of 1.8-2.6.

     As a research tool the approach of generating liquid drops and
treating them thermally appears useful, but larger, more continuous
aerosol production has not yet been shown to be feasible.

7    SUMMARY

     Exposure to metal or metallic compound submicron aerosols is
widespread, both to workers in an industrial setting and to the general
population.  In fact, the metals industries are second only to power

                                       31

-------
plants in total nationwide man-made emissions to the atmosphere.  Research
in areas of effects and control of metal aerosols requires sources of
test aerosols. Submicron aerosols are needed for inhalation experiments
(deposition and clearance of particles  in  the lungs),  air pollution
control equipment testing, filter testing,  and  sampling  instrumentation
development and calibration.  To be useful, however, the aerosol must  be
reproducible, with a predictable characteristic particle size  and  range
of diameter. Moreover, to simulate industrial emissions  of metal aerosols,
the appropriate range of diameters is from  0.25 to  5.5 ym.

     There are many possible methods for the generation  of laboratory
test aerosols.  These methods can be on a  continuous or  a batch basis,
with  high or  low concentrations and generation  rates,  and with monodisperse
or  polydisperse size distributions.  The method to  be  chosen and the
resulting aerosol depend on the requirements of the specific research
being  conducted.

      Metal aerosol generation can be thought of as  being in two major
divisions, condensation and dispersion.  Condensation  includes the
formation of  new particles from a vapor state which has  been created by
evaporation,  sublimation, or combustion of some parent material.   Dispersion
includes any  method  involving a redistribution  of the  parent material
into  a smaller particle size.  This could  include grinding for solids  or
some  droplet  formation mechanism for liquids and solids  in a liquid
suspension.   Treatment of the drops after  formation by application of
heat  for evaporation or thermal decomposition is sometimes utilized to
alter the droplet characteristics into  a more useful aerosol.   Condensation
aerosols are  usually  <1 ym, mainly existing as agglomerate chains  (0.4-
0.8 ym typical  sizes) composed of charged  primary spherical particles
 (0.001-0.01 ym).  Dispersion aerosols are  usually > 1  ym, but  are  more
controllable  and generally more monodisperse than those  formed by  condensation.

      The exploding wire generator has been studied  and used for many
years. With carefully regulated operation,  a useful reproducible metal
aerosol  can be produced, but only in small  amounts  (the  mass of the
                                        32

-------
wire is the limiting factor) and on a batch basis.  A high coagulation
rate and polydisperse aerosol are due to the local high concentration in
the plasma following an explosion.  Immediate dilution partially alleviates
this problem.

     The low intensity electric arc generator includes easily accessible
components, but offers a low metal to aerosol conversion efficiency.  The
process, consisting of vaporizing and melting the metal electrodes, is
not a highly controllable system due to the arc flame instability.  The
product is normally a polydisperse agglomerated aerosol whose reproducibility
is not yet totally proven.  It does, however, come closest to simulating
industrial emissions; from that standpoint and because of its ease of
operation, it is a worthwhile system.  Generation rates as high as 8
g/minute have been reported.  The high intensity arc offers a much
higher conversion efficiency and vaporization rate and operates with a
more controllable stable flame.  Its aerosols are still somewhat polydisperse
due to coagulation.  Electrode transpiration dilution helps reduce the
tendency to coagulate.  Some difficulty has been noted in feeding the
powderized material to the electric arc.  It has been suggested that it
may be most useful to utilize this process in the basic configuration of
an arc struck across an electrode and a metal-filled crucible.  Generation
rates as high as 40 g/minute have been reported with the high intensity
arc.

     Heating metallic material to evaporation by passing current through
the material or by direct heating in a furnace provides very small
quantities of aerosol.  Most experiments using this process are theoretical,
investigating basic principles of evaporation and condensation. Since
low concentrations are produced, coagulation is not as great a problem.
Some difficulties occur when heating wires, since the constantly decreasing
diameter creates a non-steady state current.

     Combustion processes such as spray solvent burning have high generation
rate capabilities.  The exhaust gas stream containing the aerosol may be
hot and contain products of combustion such as CO, NO , and H~0 which
                                                     A       £
may be used to simulate exhaust gases found in the industrial source

                                      33

-------
under investigation.  Combustion processes  provide particles in the 0.01
to 10 ym range which simulate closely  the range  found  in industrial
sources.

     Direct production of  solid particles from grinding or  redispersion
methods is generally not too successful  from the standpoint of  uniform
size and shape.  Liquid droplet formation methods such as  spray nebulization,
spinning disks,  and ultrasonic nebulization have been  very successful
producing monodisperse aerosols.   Generally, however,  these aerosols  are >
1  ym.   Researchers  have not yet developed this into  a  continuous,  high
concentration method.

      In summary, the  ability to generate a  known reproducible  source  of
aerosols,  specifically metallic, for laboratory  and  research purposes is
very important.   A  variety of  significant work has been conducted  in
this field.   Of  the techniques  studied only electric arcs  and  combustion
 have demonstrated the capability for continuous  generation  of  large
 quantitities  of  metal  aerosols which simulate the particle size of
 industrial  sources.
                                      34

-------
8    REFERENCES

     1.  U. S. Environmental Protection Agency, Background Information for
Standards of Performance:  Electric Arc Furnaces in the Steel Industry,
Voj^l:  Proposed Standards. EPA-450/2-74-017a (NTIS No. PB 237840),
October 1974.

     2.   Bates, C. E., and Scheel, L. D., "Processing Emissions and
Occupational Health in the Ferrous Foundry Industry,"  AIHA J^, _35_(8):452-
462, August 1974.

     3.   Brief, R. S., Rose, A. J., Jr., and Stephan, D. G., "Properties
and Control of Electric-Arc Steel Furnace Fumes,"  JAPCA. 6(4):220-224.
1957.

     4.   U. S. Environmental Protection Agency, Compilation of Air
Pollutant Emission Factors. AP-42 (NTIS No.  PB 223996), 1974.

     5.   Nerkervis, R. J., and Hallowell, J. B., "Metals Mining and
Milling Process Profiles with Environmental  Aspects,"  EPA-600/2-76-167
(NTIS No. PB 256394/AS), June 1976.

     6.   Calvert, S., et al., "Fine Particle Scrubber Performance
Tests," EPA-650/2-74-093 (NTIS No. PB 240325), October 1974.

     7.   Calvert, S., et al., "National Dust Collector Model 850 Variable
Rod Module Venturi Scrubber Evaluation." EPA-600/2-76-282 (NTIS No. PB
263617), December  1976.

     8.   Cooper,  D. W., "Pentapure Impinger Evaluation." EPA-650/2-75-
024a (NTIS No. PB  240397), March 1975.

     9.   McCain,  J. D., "Evaluation of Aronetics Two Phase Jet Scrubber,"
EPA-650/2-74-129 (NTIS No. PB 239422). December 1974.

     10.  Drehmel, D. C., "Field Test of a Venturi Scrubber in Russia."
In Second EPA Fine Particle Symposium. New Orleans, La., May 1977. EPA-
600/2-77-193  (NTIS No. PB 273828), September 1977.

                                       35

-------
     11.   Hanna, T.  R., and Pilat, 0. J., "Size Distribution of Participates
Emitted from a Horizontal Spike Soderberg Aluminum Reduction Cell,"
JAPCA, 22:533-536. 1972.

     12.   Harris, D. B., and Drehmel, D. C., "Fractional Efficiency of
Metal Fume Control as Determined by Brink Impactor," Paper 73-324 presented
at the 66th meeting of APCA, Chicago, Illinois, June 24-28, 1973.

     13.  Kolderup, H.,  "Particle Size Distribution of  Fumes  Formed by
Ferrosilicon  Production,"  JAPCA, 27.(2): 127-130, 1977.

     14.  McCord, C. P., Harrold, 6. C., and Meek, S. F.,  "A  Chemical
and  Physiological Investigation of  Electric Arc Welding,"  J.  Indust. Hyg.
and  Toxicol..  23:200-215,  1941.

      15.  Pfefferkorn,  G., and Desler, H.,  "Investigating  Smokes Generated
 in  Electric-Arc Welding,"  Staub  (English Translation),  27^(8) :l-3, 1967.

      16.  Stettler,  L.  E., Groth, D. H., and Mackay, G.  R.,  "Identification
 of  Stainless  Steel  Welding Fume Particulates in Human Lung and Environmental
 Samples  Using Electron Probe Microanalysis,"   AIHA J. 38:76-82,  1977.

      17.  Sentz, F.  C., and Radow,  A. B.,  "Exposure to  Iron  Oxide  Fume
 at  Arcair and Powder-Burning Operations,"   AIHA J. 30(2):143-146,  1969.

      18.  Hewitt, P. J., "Deposition and Elimination of Iron Oxide
 Aerosol  from  the Lung  of Rats:  Comparison  with ICRP Predictions  for
 Man,"   Proceedings  of  the Third International  Congress  of the International
 Radiation Protection Association,  September 9-14,  1973, Wash., D.  C.,
 U.  S.  AEC,  W. S. Snyder, Ed.,  Feb., 1974,  pp.  1249-1254.

      19  Casarett, L. J., and Epstein  B.,  "Deposition  and Fate of
 Inhaled l'ron-59 Oxide  in Rats,"   AIHA  J. 27:533538,  1966.

      20.  Fuchs, N. A., The Mechanics  of Aerosols,  Pergamon Press, Inc.,
 New York, 1964, pp. 1-2.
                                       36

-------
     21.  Mercer, T. T., Aerosol Technology in Hazard Evaluation,
Academic Press, 1973, 394 pp.

     22   Fuchs, N. A., and Sutugin, A. G., "Generation and Use of
Monodisperse Aerosols,"  Chap. 1, Aerosol Science, C. N. Davies (ed.),
Academic Press, Inc., New York, 1966.

     23.  Corn, M., and Esmer, N. A., "Aerosol Generation," Chap. 2,
Handbook on Aerosols, Richard Dennis (ed.), U. S. ERDA, TID-26608, 1976.

     24.  Chase, W. G., and Moore, H. K. (eds.), Exploding Wires. Vols.
1-4, Plenum Press, New York, 1958-1968.

     25.  Karioris, F. G., and Fish, B. R., "An Exploding Wire Aerosol
Generator." J. Colloid Science, 1_7_:155-161, 1962.

     26.  Duerksen, K. D., Walker, R. L., and Karioris, F. G., "Exploding-
Wire Aerosol Generator,"  In:  Health Physics Division Annual Progress
Report for Period  Ending July 31, 1967, Oak Ridge National Lab., Tenn.,
Contract W-7405-ENG-26, pp. 296-298, 1967, ORNL-4168.

     27.  Smith, B. M., Wagmen, J., and Fish, B. R., "Interaction of
Airborne Particles with Gases,"  Environmental Science and Technology,
1(6):558-562,  1969.

     28.  Tomaides, M., and Whitby, K. T., "An Improved Exploding Wire
Aerosol Generator,"  Preprint, Minnesota Univ., Minneapolis, 6 pp.,
1969.  EPA/APT1C No. 19409.

     29.  Phalen,  R. F.,  "Evaluation of an Exploding-Wire Aerosol Generator
for Use in Inhalation Studies,"  J. Aerosol Science, 3.:395-406, 1972.

     30.  Sherman, P. M., "Generation of Submicron Metal Particles," JL_
Colloid and Interface Science, J51_:87-93, 1975.

     31.  Wegrzyn, J. E., "An Investigation of an Exploding Wire Aerosol,"
In Proceedings: Symposium on Fine Particles:  Minneapolis, Minn.. May
1975. EPA-600/2-75-059  (NTIS No. PB 249514),  pp. 253-273, October 1975.

                                      37

-------
     32.   Kops, J., VandeVate, J. F., Jermans, L., and Dibbets, G., "The
Aerodynamic Diameter of Branched Chain-like Aggregates,"  Nature.
252:385-387, 1974.

     33.   Kops, J., Dibbets, G., Hermans, L., and VandeVate, J. F., "The
Aerodynamic Diameter of Branched Chain-like Aggregates,"  J. Aerosol Science,
6:329-333, 1975.

     34.  Whytlaw-Gray, R., Speakman, J. B., and Campbell, J.  H.  P.,
"Smokes:  Part  I  - A Study of Their Behaviour and a Method of  Determining
the Number of Particles They Contain,"  Proceedings of the Royal  Society
of London. Series A, Vol. 120, pp. 600-615, 1923.

     35.  Amson,  J. C., "Electric Voltage in the Consumable-Electrode
Arc System,"  J.  Physics D:  Applied Physics. 5;89-96, 1972.

     36.  Carslaw, H. S., and Jaeger, J. C., Conduction of Heat in
Solids,  2nd ed.,  London:  Oxford University Press, 1959.

     37.  Homma,  K., Koshi, S., and Sakabe, H., "Study on the  Preparation
Methods  of  Submicron Particles,"  Bull. Nat. Inst. Indust. Health.  2_:57-
62,  1959.

     38.  Harvey,  J., Matthews, H. I., and Wilman, H., "Crystal Structure
and Growth  of Metallic  or Metallic-oxide Smoke Particles Produced by
Electric Arcs,  "   Discussion  of the Faraday Society,  30:113-124,  1960.

     39.  Reichelt, H.,  "On Aerosols Generated in an  Electric  Arc Between
Metal  Electrodes,"  In:  Aerosol Research at the First Physics Institute.
Vienna University, Austria, pp. 19-29, Jan.. 1968.

     40.   Linsky, B., Hedden, R., Naylor, M., and Dimmick F.,  "Generation
and  Simulation  of Metallic  Particulate Air Pollutants by Arc Metal  Wire
Spraying,"  EPA-600/2-77-208  (In Print), October 1977.

     41.  Zimmerman, N., EPA/IERIL-RTP, Particulate Technology  Branch,
unpublished data, 1977.

                                        38

-------
     42.  Sheer, C., and Korman, S., "The High-Intensity Arc in Process
Chemistry,"  In:  Arcs in Inert Atmospheres and Vacuum, W. E. Kuhn, ed.,
John Wiley and Sons, N. Y., pp. 169-183, 1956.

     43.  Harris, V., Holmgren, J. D., Lorman, S., and Sheer, C., "Arc
Decomposition of Rhodonite,"  J. Electrochemical Society. 106: 874-876,
1959.

     44.  Holmgren, J. D., Gibson, J. 0., and Sheer, C., "Some Characteristics
of Arc Vaporized Submicron Particulates,"  J. Electrochemical Society,
111: 362-369, 1964.

     45.  Korman, S., Sheer, C., Angier, D. J., and Shaw, H., "Properties
of Submicron Oxides Produced by Arc Vaporization,"  International  Conf. on
Fine Particles. 1973. pp. 153-164, The Electrochemical Society, Inc.,
1974.

     46.  Korman, S., Personal Communication, Columbia University,
 New  York, April 1977.

     47.  Boffa, C. V., and Pfender, E., "Controlled Generation of
Monodisperse Aerosols in the Submicron Range,"  J. Aerosol Science,
4:103-112, 1973.

     48.  Boffa, C. V., Mazza, A., and Rosso, D. M., "Generation of
Monodisperse Submicron Aerosols by Ablation from Transpiration-Cooled
Porous Matrices," In Proceedings:   Symposium on Fine Particles:  Minneapolis.
Minn., May 1975. EPA-600/2-75-059 (NTIS No. PB 249514), pp.  111-127,
October 1975.

     49.  Tarroni, G., Prodi, V.,  Melandri, C., Mompane, G.  F., DeZaiacomo,
T., and Formignani, M., "Production of Ultrafine Monodisperse Aerosols
by Condensation,"  J. Aerosol Science, 6:305-310, 1975.

     50.  Fannick, N. F., and Corn, M., "The Plasma Jets:  Industrial
Hygiene Aspects and a Survey of Current United States Practices for
Employee Protection," AIHA J.. 30(3):226-235, 1969.

                                       39

-------
     51.  Turkevich, J.f "UHrafine Particles in the Gas Phase,"  Fundamental
Phenomena in the Material Sciences, 3_:195-212, 1966.

     52.  Spumy, K., and Lodge, J. P., Jr., "Radioactively Labelled
Aerosols,"  Atmospheric Environment. 2;429-440, 1968.

     53.  Spurny, K., and Jampl, V., "Preparation of Radioactively
Labelled Condensation Aerosols.  I - Aerosols of Sodium Chloride, Silver
Iodide and Sulphuric Acid,"  Collection Czechoslav. Chem. Comtnun.,
30:507-514, 1965.

     54.  Spurny, K., and Lodge, J. P., Jr., "Preparation of Radioactively
Labelled Condensation Aerosols.  IV - Aerosols of Gold, Rhenium Oxide,
Silver, Tellurium, and Vanadium Oxide,"  Collection Czechoslav. Chem.
Commun.. 36_: 3358-3362, 1971.

     55.  Spurny, K., "Preparation of Radioactively Labelled Condensation
Aerosols,"  AIHA J.. 33:431-432, 1972.

     56.  Spurny, K., and Lodge, J. P., Jr., "The Generation of Highly
Dispersed Radioactive Labelled Aerosols,"  Presented at the Tenth Colloquium
on Atmospheric Pollution (Dixieme Colloque Sur Les Atmospheres Polluees),
Paris, May 3-5, 1975.

     57.  Spurny, .K., and Lodge, J. P., Jr., "Production of Highly
Dispersed Model Aerosols for Dust Research and Filter Testing,"  Staub
(English Translation), 33_(4):171-174, 1973.

     58.  Heicklen, J., and Luria, M., "Kinetics of Homogeneous Particle
Nucleation and Growth,"  International J. Chemical Kinetics. (Symp. 1,
1974), 7.:567-580, 1975.

     59.  Maiwald, E., "A Tungsten Oxide Test Aerosol,"  Staub (English
Translation), 25_(12):13-15, 1965.

     60.  Polydorova, M., "The preparation and Properties of a Tungsten
Oxide Aerosol,"  Staub (English Translation), 25(12):16-18, 1965.
                                       40

-------
     61.  Polydorova, M., "The Manufacture and Properties of a Platinum
Oxide Aerosol,"  Staub (English Translation), 27_(8)-5-8, 1967.

     62.  Tomb, T. F., Treaftis, ri, M., and Beckert, A. F., "Quantitating
the Efficiency of the Coal Mine Personnel Respirable Dust Sampler for
Sampling Oxide Fumes,"  AIH/Lk' 37:459-463, 1976.

     63.  Barchet, W. R., and McKenzie, R. S., "Silver Iodide Aerosol
Production by a Plated Hot-Wire Generate"","  -I^Jtogligd..Meteorology,
1^:1151-1155, 1975.

     64.  Poc, M., and Roulleau, M., "The Ice Nucleating Properties of
an Ultrafine Aerosol  of Pure Silver Iodide,"  J. Applied Meteorology,
14:1146-1150, 1975.

     65.  Buckle, E.  R., and Poincon, K. C., "Condensation and Evaporation
of Metallic Aerosols,"  Faraday Symposia of the Chemical Society, 3:78-84,
1973.

     66.  Buckle, E.  R., and Pointon, K, C., "Growth and Sedimentation in
Cadmium Aerosols,"  Atmospheric Environment, 3:1335-1337, 1974.

     67.  Buckle, E.  R., and Pointon, K. C., "Condensation of Cadmium
Aerosols,"  J. Materials Science,  10:365-378, 1975.

     68.  Billings, C. E., Small,  V<. 3., and Silver-man, I., "Pilot-Plant
Studies of a Continuous Slag-Wool  Filter for Open Hearth Fume,"  JAPCA,
^(3):159-165, 1955.

     69.  Brain, J. D., Valberg, P. A.,  Sorokin, S. P., and Hinds, W. C.,
"An Iron Oxide Aerosol Suitable for Animal  Exposures,"  Environmental
Research. 7:13-26, 1974.

     70.  Glauberman, H., and Breslin, A. J., "The Generation of a Solid
Sub-micron Test Aerosol,"  AIHA Quarterly, ]_7:401-404,  Dec.,  1956.
                                       41

-------
     71.  Formenti, M., Juillet, F., Meriaudeau, P., Teichner, S. J.,
and Vergnon, P., "Preparation in a Hydrogen-Oxygen Flame of Ultrafine
Metal Oxide Particles,"  J. Colloid Interface Science. 39:79-89, 1972.

     72.  Juillet, F., LeComte, F., Mozzanega, H., Teichner, S. J.,
Thevenet, A., and Vergnon, P., "Inorganic Oxide Aerosols of Controlled
Submicronic Dimensions,"  Faraday Symposia Chemical Society, ,3:57-62,
1973.

     73.  Dharmarajan, V., and West, P. W., "A Precise Method for the
Generation of Standard Metal Salt Particulates,"  Analytica Chimica
Acta. 69:43-48, 1974.

     74.  Carroz, J. W., Odencrantz, F. K., and Finnegan, W. G., "Generation
of  Fumes  Simulating Particulate Air Pollutants," EPA-600/2-77-132,
July 1977.

     75.  Albert, R. E., Petrow, H. G., Salam, A. S., and Spiegelman,
J.  R.,  "Fabrication of Monodisperse Lucite and Iron Oxide Particles with
a Spinning  Disk Generator,"  Health Physics, 10:933-940, 1964.

     76.  Spertell, R. B., and Lippman, M., "Airborne Density of Ferric
Oxide Aggregate Microspheres,"  AIHA J. 32(11):734-740, 1971.

     77.  Caplan,  K. J., Doemeny, L. J., and  Sorenson, S. D., "Performance
Characteristics of  the 10 mm Cyclone Respirable Mass Sampler:  Part 1 -
Monodisperse  Studies,"  AIHA J. 38:83-95, 1977.

     78.  Waite,  D. A., and Ramsden, D., "The Production of Experimentally
Labelled  Aerosols  in the Sub-Micron Range,"   J. Aerosol Science, 2;425-
436, 1971.

     79.  Raabe,  0. G., Kanapilly, G.  M., and Newton, G. J.,  "New Methods
for the Generation  of Aerosols of Insoluble Particles for Use in  Inhalation
Studies,"  In:  Inhaled Particles III  - Proceedings of an International
Symposium.  Sept.  14-23, 1970, London, Vol. 1, pp.  3-18, 1971.
                                      42

-------
     80.  Kotrappa, P., and Wilkinson, C. 0., "Densities in Relation to
Size of Spherical Aerosols Produced by Nebulization and Heat Degradation,"
AIHA J., 33:449-453, 1972.

     81.  Horstman, S., Barkley, W., Larson, E., and Bingham, E., "Aerosols
of Lead, Nickel and Cadmium,"  Archives of Environmental Health, 2i5:75-
77, 1973.

     82.  Friedman, J. M., and Horstman, S. W., "A Simple Method for the
Generation of Metallic Oxide Aerosols,"  AIHI J, 35825-831, 1974.
                                      43

-------
                                TECHNICAL REPORT DATA
                         (Please read Inductions on the reverse before completing)
 . REPORT NO.
  EPA-600/7-78-013
                           2.
                                                      3. RECIPIENT'S ACCESSION NO.
 . TITLE AND SUBTITLE
 Characterization and Generation of Metal Aerosols
                                                       5. REPORT DATE
                                                        February 1978
                                                       6. PERFORMING ORGANIZATION CODE
 .AUTHOR(S)Neil Zimmermari) Dennis C.  Drehmel, and
           James H. Abbott
                                                       8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 U.S. Environmental Protection Agency
 Industrial Environmental Research Laboratory
 Research Triangle Park, North Carolina 27711
                                                       10. PROGRAM ELEMENT NO.
                                                      EHE623
                                                      11. CONTRACT/GRANT NO.

                                                      NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC  27711
                                                      13. TYPE OF REPORT AND PERIOD COVERED
                                                      Final: 6/76-6/77    	
                                                      14. SPONSORING AGENCY CODE
                                                        EPA/600/13
15. SUPPLEMENTARY NOTES
61, 919/541-2925.
                              project off icer/auttior is D.C. Drehmel, Mail Drop
 16 ABSTRACT The report reviews techniques of metal aerosol generation for the purpose
 of establishing the state-of-the-art of the technology and guiding future researchers.
 Exposure to metal or metallic compound submicron aerosols is widespread in both
 industrial and general environments.  Research in areas of health effects, sampling
 instrumentation,  and air pollution control technology requires a reliable source of
 test aerosol.  The report presents many mechanisms for generating metal aerosols,
 and discusses their applications, advantages, and disadvantages. Generation  methods
 can be on a continuous or batch basis, with high or low concentrations and generation
 rates, and with monodisperse or polydisperse size distributions. The  method chosen
 and the resulting aerosol depend on the requirements  of the specific research being
 conducted.   Metal aerosols generated by exploding wires, electric arcs, heating to
 evaporation, combustion, and dispersion are presented, with particular attention paid
 to particle  size characteristics.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                          b.IDENTIFIERS/OPEN ENDED TERMS
                                                                   c.  COSATI Field/Group
 Air Pollution
 Aerosols
 Metallurgy
 Aerosol Generators
 Properties
 Testing
                      Exploding Wires
                      Electric Arcs
                      Evaporation
                      Combustion
                      Dispersions
                      Size Determination
Air Pollution Control
Stationary Sources
Particulates
Metal Aerosols
Characterization
13B
07D
11F
15B
14B
19A,14B
  20C

  21B
18. DISTRIBUTION STATEMENT
 Unlimited
                                           19. SECURITY CLASS (This Report)
                                           Unclassified
                                                                   21. NO. OF PAGES
                                                                         50
                                          20. SECURITY CLASS (Thispage)
                                          Unclassified
                                                                    22. PRICE
EPA Form 2220-1 (9-73)
                                        44

-------