EPA-650/2-74-027
MAY  1974
Environmental Protection Technology  Series
                                           Wi£  o
                                                             UJ
                                         mM^M^^^^^M
                                         ^^SS^^^^^X^i^


-------
                            EPA-650/2-74-027
CONTROL  TECHNOLOGY
FOR  FINE  PARTICULATE
        EMISSIONS
               by

           L. J. Shannon

       Midwest Research Institute
         425 Volker Boulevard
      Kansas City, Missouri  64110
         Grant No. R-801615
         ROAPNo. 1AB012-05
      Program Element No. 1AB012
   EPA Project Ofliccr:  J. M. Shackclford
      Air Pollution Control Divjsion


           Prepared for

  OFFICE OF RESEARCH AND DEVELOPMENT
 U.S. ENVIRONMENTAL PROTECTION AGENCY
       WASHINGTON, D.C. 20460

             May 1974

-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication.  Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
                                   ii

-------
                              ABSTRACT

This report gives results of a study to assess the state of the art of
control technology for fine particulates emitted from stationary sources.
It emphasizes the analysis of control technology for primary particulates.
The initial activity was a general review of collection and agglomera-
tion mechanisms for particulates.  Both theoretical and experimental facets
were reviewed with attention on forces or collection mechanisms expected
to be of importance for fine particles.  The analysis of control tech-
nology for fine particulates centered on conventional control equipment,
on emerging control technology, and on proposed or conceptual control
systems.
                                   iii

-------
                          TABLE OF CONTENTS

                                                                 Page

Abstract	ill

List of Figures	ix

List of Tables	xii

Acknowledgements	xiii

Conclusions 	  1

  Particle Collection and Agglomeration Mechanisms	1

    High Velocity Differential Systems	3
    Low Velocity Differential Systems 	  4
    Electrostatic Forces	4
    Diffusiophoretic Forces 	  4

  Agglomeration and/or Growth of Particles	5
  Conventional Control Equipment	5
  Emerging and Proposed Control Systems 	  6

Recommendations for R&D Programs	8

  Collection and Agglomeration Mechanisms 	  8

    Studies of the Dynamics of Particulate Agglomeration. ...  8
    Sonic Agglomeration 	  9
    Theoretical and Experimental Investigation of the
      Dynamics of Mixing of Steam, Water Droplets and
      Particulates	9
    Particle Charging 	 10

  Conventional Control Equipment	10

-------
                     TABLE OF CONTENTS (Continued)
    Electrostatic Frecipitators	11
    Fabric Filters 	  11
    Wet Scrubbers	11

  Emerging Control Systems 	  12
  Proposed Control Systems 	  12

    Condensation Scrubbers 	  12
    Charged Droplet Systems	13
    Electrified Filters	13
    Foam Scrubbers 	  13

Introduction 	  14

Collection and Agglomeration Mechanisms for Farticulates ....  16

  Particle Collection	16

    Gravity and Momentum Forces	16
    Centrifugal Forces 	  18
    Aerodynamic Capture	18
    Flux Forces	19

  Particle Agglomeration and Growth	22

    Thermal Agglomeration	22
    Turbulent Agglomeration	22
    Sonic Agglomeration	23
    Agglomeration of Charged Particles 	  24
    Particle Growth by Condensation	25

  Summary	26

Fine Particle Collection by Conventional Control Equipment ...  27

  Electrostatic Precipitators	28
  Fabric Filters 	  33
  Wet Scrubbers	35
  Summary	36
                                    vi

-------
                    TABLE OF CONTENTS (Continued)

                                                                 Page

Emerging and Proposed Control Systems for Fine Particulates.  .  .   38

  Emerging Control Systems 	   38

    Steam-Hydro Air Cleaning System	40
    ADTEC Wet Scrubber 	   41
    Wetted-Knit Mesh Filters 	   42

  Proposed Control Systems 	   43

    Condensation Scrubbers 	   43
    Charged Droplet Scrubbers	47
    Foam Scrubbers	48
    Electrified Filters	49
    Granular Bed Filters	49
    Fluidized Beds	50
    Gamma-Ray Precipitator 	   50
    Sonic Agglomerators	51

  Summary	51

References	53

Glossary of Terms, Abbreviations, and Symbols	56

Appendix A - Collection and Agglomeration Mechanisms for Fine
               Particulates	59

Appendix B - Emerging Control Technology—Specific Control
               Devices	100

Appendix C - Condensation Scrubbing for Particulate Collection .  132

Appendix D - Charged Droplet Scrubbers as Particulate Control
               Devices	151

Appendix E - Foam Scrubbers as Particulate Control Devices .  .  .  161
                                   vii

-------
                    TABLE OF CONTENT (Concluded)
Appendix F - Electrified Filters as Farticulate Control
               Devices	168

Appendix G - Granular Bed Filters as Particulate Control
               Devices	175

Appendix H - Fluid Beds as Particulate Control Devices	188

Appendix I - Gamma-Ray Precipitators as Particulate Control
               Devices	203
                                   viii

-------
                           LIST OF FIGURES

No.                                                              Page

1    Measured and computed efficiency as a function of particle
       size for electrostatic precipitator installation at a
       coal-fired power plant	31
2    Variation of electrostatic precipitator efficiency with
       size of particle dust (sizing by electron microscope) . .  32
3    Measured efficiency of Dupont wetted-knit mesh separator
       collecting 2 micron particles	44
A-l  Ratio of particle velocity to collector body diameter
       necessary to achieve 90% target efficiency for inertial
       impact ion	65
A-2  Particle velocities required to achieve 90% target effi-
       ciency for inertial impaction 	  66
A-3  Interception parameter as a function of Reynolds Number
       for selected target efficiencies	68
A-4  Target efficiency by diffusion on an isolated cylinder. . .  70
A-5  Approximate solutions of equations for target efficiences .  72
A-6  Calculated particle migration velocities for submicron
       particles	74
A-7  The thermophoretic velocity in air as a function of tem-
       perature gradient	77
A-8  The diffusiophoretic velocity in helium as a function of
       water-vapor pressure gradient at temperatures between
       20° and 50°C	80
A-9  Particle agglomeration flux for various particle
       diameters and fluid velocities	87
A-10 Coagulation constants for various particle diameters and
       fluid velocities	87
A-11 Comparison of theoretical predictions with Yoder and
       Silveman's agglomeration data	88
A-12 Effect of particle number concentration on agglomeration
       rate	94
                                    ix

-------
                     LIST OF FIGURES (Continued)

No.                                                              Page

A-13  Effect of particle mass concentration on agglomeration
        rate	   94
A-14  Effect of particle charge on agglomeration	   95
B-l   Simplified diagram of Steam-Hydro Air Cleaning System .  .  102
B-2   Diagram of steam nozzle and water spray nozzles (Steam-
        Hydro Air Cleaning System)	103
B-3   Diagram of one Steam-Hydro air cleaning unit serving
        open-hearth furnace 	  104
B-4   Schematic diagram of basic Aronetics System 	  107
B-5   Mixing section pressure rise for ADTEC scrubber (1 kg
        H20/kg gas)	109
B-6   Schematic diagram of nucleation scrubber	115
B-7   Schematic diagram of installation of nucleation scrubber
        systems on fiberglass furnace 	  117
B-8   Single-stage dynactor system (cross-sectional view) . .  .  120
B-9   Schematic diagram of Pentapure™ Impinger	122
B-10  Schematic diagram of the Mystaire™ laboratory scrubber.  .  124
B-ll  Schematic diagram of Dupont wetted-fiber bed scrubber .  .  126
B-12  Measured collection efficiency of Dupont wetted-knit
        mesh separator (2 micron particles)	128
B-13  Correlation of flooding velocity of Dupont scrubber . .  .  129
C-l   Cumulative mass fraction distribution of experimental
        aerosols used by Prakash and Murray	140
C-2   Five-stage Peabody perforated plate scrubber adapted for
        steam injection	143
C-3   Five-stage steam injection plate scrubber 	  144
D-l   Classification of particle-drop interaction mechanisms
        (Melcher and Sachar)	152
D-2   Experimental apparatus for space-charge precipitation .  .  156
E-l   The Centrifoam Scrubber	165
F-l   Collection efficiency of wire grid filters with coal
        dust	170
F-2   Collection efficiency of wire grid filters with
        quartz dust	170
F-3   Penetration load curve for electret filter	172
G-l   Electrified packed bed	183
G-2   Performance characteristics of electrostatically
        agumented packed bed	184

-------
                      LIST OF FIGURES (Concluded)

No.                                                              Page

H-l   Schematic flow diagram of Black's fluid bed system. ...  190
H-2   Results of pressure drop measurements across fluidized
        bed at various gas flow rates and bed heights 	  192
H-3   Collection efficiencies in fluidized beds 	  194
H-4   Collection of 0.67 um OOP aerosol by beds of 175 urn
        alumina granules	196
H-5   Possible configuration for electrofluidized bed, typical
        of application to cleaning of oil ash	199
1-1   Reactions occurring in the radiation-induced pre-
        cipitator	204
1-2   Space charge separation and particle charging in radia-
        tion-induced precipitator 	  204
1-3   Schematic diagram of gamma-ray precipitator and aux-
        iliary equipment at Pennsylvania State University .  . .  206
1-4   Collection efficiency of gamma-ray precipitator vs
        central electrode potential based on weight percent
        of inlet particulate concentration	208
1-5   Collection efficiency of gamma-ray precipitator vs
        dose rate	208
1-6   Experimental fractional efficiencies of gamma-ray pre-
        cipitator 	209
                                   xi

-------
                           LIST OF TABLES

No.                                                              Pag

1     Promising approaches for collection or agglomeration
        of fine particles	2
2     Particle collection or agglomeration forces (mechanisms) . 17
3     Estimated fractional efficiencies of electrostatic pre-
        cipitator	29
4     Emerging and proposed control systems for fine par-
        ticulates	39
5     Test data on Dupont scrubber	45
A-l   Particle collection or agglomeration forces (mechanisms) . 60
A-2   Equations' for thermophoretic velocity	76
B-l   Data summary for Chromasco pilot model ADTEC scrubber. . .110
B-2   Particle size distribution of ferro-silicon dust 	112
B-3   Performance tests on Teller nucleation scrubber (con-
        tainer glass furnace)	116
B-4   Performance tests on Teller nucleation scrubber (fiber-
        glass furnace) 	116
B-5   Test data on Dupont scrubber 	127
C-l   Summary of selected experimental investigations of steam
        condensation	136
D-l   Iron oxide data, no fog added to the precipitator	157
D-2   Iron oxide data, with fog added to the precipitator. . . .158
6-1   Summary of selected experimental investigations of
        granular bed filters 	178
H-l   Estimated collection by multiple-contact fluidized
        beds	197
1-1   Typical particle concentration at inlet to gamma-ray
        precipitator (20-ft3 sample)	207
                                   xii

-------
                          ACKNOWLEDGEMENTS

This report was prepared for the Air Pollution Control Division, Office
of Research and Development under Grant No. R-801615.  The work was per-
formed in the Environmental Systems Section of'the Physical Sciences Divi-
sion.  The report was written by Dr. L. J. Shannon, Head, Environmental
Systems Section.  Dr. A. E. Vandegrift, Dr. Frank T. Greene, and Mr. Paul
Gorman contributed to several phases of the analyses discussed in the
report.

Several people outside Midwest Research Institute contributed informa-
tion that was used in the preparation of this report.  The program could
not have been completed without the assistance of EPA/OR&D personnel who
graciously assisted MRI in establishing contacts with various groups
actively working on fine particulate control.  Foremost in this regard
were Mr. Richard Harrington, Dr. Arnold Goldberg, Mr. Kurt Yeager, and
Dr. James Shackelford of the Air Pollution Control Division and Mr. Jim
Abbott, Mr. Alfred Craig, Mr. Robert Lorentz, Dr. Les Sparks, Dr. Dennis
Drehmel and Mr. Dale Harmon of the Control Systems Laboratory.

Special thanks are also given to research groups at Southern Research
Institute, Stanford Research Institute, APT., Inc., and the University
of Washington who permitted MRI personnel to visit their laboratories
and discuss in detail their research activities.
Approved for:

MIDWEST RESEARCH INSTITUTE
tjjV
H. M. Hubbard, Director
Physical Sciences Division
                                  xiii

-------
                             CONCLUSIONS

Presently available equipment for the control of particulate emissions
from stationary sources has achieved limited success in the control of
fine particulate pollutants (i.e., less than or equal to 3 micron diam-
eter particles).  The inability of existing control equipment to collect
the fine particulates means that if we are to achieve significant control
in this area an aggressive, well-conceived research and development pro-
gram will be required to improve existing equipment and develop new tech-
niques for the collection of fine particulates.

In order to assist EPA in formulating an overall R&D program on control
technology for fine particulates, Midwest Research Institute (MRI)  con-
ducted an assessment of the current and future potential for the con-
trol of fine particulates under the auspices of the Air Pollution Control
Division, Office of Research and Development, Environmental Protection
Agency.  The major areas included in the evaluation were:

1.  Particle collection and agglomeration mechanisms or forces that may
be effective in the submicron size range;

2.  Conventional control equipment for particulates; and

3.  Emerging or proposed control devices that appear promising, but
have not been reduced to practice on a commercial scale.

The principal results of our assessment are presented in the following
subsections.

PARTICLE COLLECTION AND AGGLOMERATION MECHANISMS

The review of collection and agglomeration mechanisms for particulates
served to pinpoint approaches and underlying mechanisms or forces which
have potential for use in the control of fine particulates.  Table 1
lists the approaches that appear to have significant potential.  While

-------
          Table 1.  PROMISING APPROACHES FOR COLLECTION OR
                   AGGLOMERATION OF FINE PARTICLES
1.  Particle collection

         A.  High velocity differential between fine particles and
               collector bodies

         B.  Low velocity differential between fine particles and
               collector bodies with both particles and collector
               bodies charged

         C.  Electrostatic and diffusiophoretic forces
II.  Agglomeration and/or particle growth

         A.  Electrostatic agglomeration

         B.  Sonic agglomeration

         C.  Condensation,

-------
the approaches delineated in Table 1 offer possibilities for use with
fine particles, realization of the potential will require the develop-
ment of viable control systems.  In some instances,  it may be possible
to modify or redesign conventional control systems to take advantage
of the promising mechanisms.  Alternatively, new control systems will
have to be developed.  Some avenues that might be pursued to utilize
the approaches suggested in Table 1 are discussed next.

High Velocity Differential Systems

Systems utilizing high velocity differentials between the particles and
collector bodies would achieve particle collection by inertial impac-
tion.  In order to achieve 90% single particle target efficiency* for
particles in the 0.1-1.0 micron particle size range, velocity differen-
tials in the range of 102-10^ m/sec and collecting bodies of diameter
50 microns or less are required.  Velocity differentials of that magnitude
can generally be achieved for less energy expenditure by imparting the
high velocity to the collector bodies rather than the gas stream.  Since
the velocity differential required to attain a given target efficiency
decreases with decreasing size of the collector body, the technique
employed to achieve the high velocity of the collector body should also
generate the smallest collection bodies possible.

Systems utilizing small, high velocity water drops as the collecting
bodies are the most practical.  Conventional venturi scrubbers operate
on this principle, but the energy requirements (i.e., pressure drop)
for fine particle collection are high.  Techniques to generate the small,
high velocity water drops at a lower energy expenditure are needed.

Nozzle(s) configurations which can both accelerate and shear the water
stream into small drops should be investigated.  The wet scrubbing sys-
tems developed by Lone Star Steel and Aronetics, Inc.  (see Appendix B)
represent new control systems that employ small, high velocity water
drops for particle collection.  Both of these systems  use waste process
heat rather than external energy to provide the energy required to
establish the requisite particle-droplet differential velocities and
move the gases through the system.
   Single particle target efficiency is defined as

    cross-sectional area of fluid stream from which particles are removed
    cross-sectional area of filter element projected in direction of flow

-------
Low Velocity Differential Systems

Low velocity differential systems of several configurations might be used
to collect fine particles—especially if electrostatic forces are utilized
to augment the other collection forces.  Electrostatic forces might be
used advantageously for cleaning gases because they can be significantly
greater than mass forces.  Fabric filters, wire mesh filters, granular
beds, or fluid beds could be used as the underlying filter elements.

Theoretical and limited experimental studies have shown that electro-
static forces can improve the collection efficiency of the filter sys-
tems mentioned in the preceding paragraph (see Appendices F, G, and H).
To date, experimental studies of electrified filters have been confined
to laboratory-scale equipment or clean, high porosity systems.  It is
not clear whether the improved performance of electrified filters noted
in the laboratory scale equipment can be realized in commercial-scale
equipment under actual industrial gas cleaning conditions.  Also, it is
not certain that any increase in collection efficiency which might be
realized in commercial-scale equipment will be sufficient to justify the
use of electrified filters.  Additional R&D will be required to answer
these questions.

Electrostatic Forces

Electrostatic forces between charged particles and collecting bodies are
quite effective for the collection of particulates.  To secure rapid
and efficient deposition of particles, forced charging of particles should
be used.  Electrostatic precipitators and various types of electrified
filters (see preceding section) can be used to take advantage of electro-
static forces for particle collection.

Particle migration velocities calculated from theoretical equations for
field and diffusion charging indicate that migration velocities in the
range of 2 cm/sec should be realized for fine particles in electrostatic
precipitators operating with reasonable electric fields, charge densi-
ties and residence times.  Therefore, significant collection of fine
particulates should be possible in electrostatic precipitators.

Diffusiophoretic Forces

The  diffusiophoretic  force, which is a combination of  forces due to Stefan
flow and a concentration gradient, should be a useful  force  to exploit  in
collection of small particles.  The  fundamental mechanism is not sensitive

-------
 to particle size and becomes more important compared to other removal
 mechanisms for particles below 2 microns.  To exploit the mechanism of
 diffusiophoresis, the collecting device must be designed such that one
 component in  the gas phase is diffusing toward a collecting surface.
 The most practical case is the diffusion of water vapor toward a sur-
 face.  Recent studies (see Appendix C) indicate that diffusiophoresis
 can produce appreciable particle deposition under the heat and mass
 transfer conditions of a realistic wet scrubber.

 AGGLOMERATION AND/OR GROWTH OF PARTICLES

 The agglomeration or growth of fine particles, as a step in the control
 of fine particulates, is in principle attractive.  If large particles
 can be produced for reasonable energy expenditures, it may be possible
 to utilize devices which cause particle agglomeration or growth in con-
 junction with conventional control systems.

Condensation is the most feasible way to achieve particle growth with-
out agglomeration.  However,  rather than utilize a separate chamber to
grow particles via condensation, incorporation of condensation effects
directly into wet-scrubber systems is recommended.

A potential way to agglomerate fine particles to a larger size is to pass
 the aerosol through a sonic field.  Sonic fields can effectively induce
agglomeration in all sizes and types of particulate materials.  However,
 the minimum particulate loading that can be treated is  about 1.1-2.3 g/m^
 (0.5-1.0 grains/ft^).  For more disperse systems it is  necessary to seed
 the gas to bring the particle loading up to the requisite level.

 The addition  of a bipolar charge to an aerosol can produce a significant
 increase in the rate of agglomeration of submicron particulates.  A bi-
 polar ly charged polydisperse aerosol containing a large number of rela-
 tively large  particulates, as well as fine particulates, appears to be
 the most favorable situation because it may be possible to collect
numerous fine particulates on the large particulates before charge anni-
hilation occurs.

 CONVENTIONAL CONTROL EQUIPMENT

Well-designed conventional control systems are capable of achieving
 relatively high collection of fine particulates.  However, more often
 the performance of conventional devices is less than satisfactory.   The
poor performance of most conventional equipment results from:

-------
1.  Inherent limitations of conventional equipment in the fine particle
size range; and

2.  Lack of previous need to control fine particulates and the resultant
lack of interest in optimizing systems with emphasis on the collection
of fine particulates.

Item 1 results from the fact that the mechanisms or forces exploited in
a specific device are not effective in the fine particle size range.
The second item, which is only applicable to control systems not in-
herently limited by the forces or collection mechanisms utilized in the
systems, is a result of previous priorities established for the control
of particulate emissions.

Although conventional control equipment is often not very effective for
collecting fine particles, certain  types of devices can be quite effec-
tive—especially  if  their performance is optimized for fine particulates.
High efficiency electrostatic precipitators offer considerable potential
in  this regard.   Changes within  the present state of the art  (i.e., in-
creases in the collection area to volume ratio, charge density, and field
strength) could produce significant improvement in the collection of fine
particulates  from some sources.  Consideration should also be given to
the further development of wet electrostatic precipitators—especially
for use with high resistivity dusts.

Improvements  in fabric filter technology might also result in better
control of fine particles from some sources.  Avenues to pursue to im-
prove  performance are not as clear-cut as with electrostatic precipitators.
Improvements  in fabric filters which would permit their use on a wider
spectrum of emission sources would  be an important advance.

Improvements  in the  performance  and reduction in power consumption of
venturi or orifice type wet scrubbers might be achieved by taking advan-
tage of condensation effects which  can occur in these types of wet
scrubbers.  Improvements in low  energy scrubbers could result from the
enhancement of  the contribution  of  flux forces  (i.e., diffusiophoresis,
thermophoresis, condensation).

EMERGING AND  PROPOSED CONTROL SYSTEMS

In  the category of emerging and  proposed control systems for  fine parti-
culates,  there  are several systems  that hold promise for  transition to
commercially  viable systems.  At the present time, foremost in  the category
of  emerging systems  are  the Steam-Hydro Air Cleaning and the  ADTEC wet-
scrubbing systems.  Both of these  systems  appear applicable to  a wide

-------
range of sources, but economics will probably limit their use to sources
where waste heat is available.  Where waste heat is not sufficient to
meet the power requirements of the systems, supplemental heat may be used.
However, the use of supplemental heat would add significantly to the
overall cost of both systems.

Only limited data from laboratory-scale systems are available to judge
the potential of most of the proposed systems.  Assessments must be based
predominantly on the ability of the underlying collection forces or
mechanisms to collect fine particulates.  Condensation scrubbers, charged
droplet systems, electrified filters of various types, and foam scrubbers
have sufficient potential to merit additional research and development.
Future studies on any of the promising proposed systems should be con-
ducted in equipment configurations and sizes that will produce reliable
engineering design data.

-------
                   RECOMMENDATIONS FOR R6D PROGRAMS

Control technology for fine particulates is at a very early stage of de-
velopment and well-conceived research and development programs are needed
to improve and develop equipment for the collection of fine particulates.
Our review identified several areas we recommend as the subjects for
additional investigation.  The Environmental Protection Agency through
its Control Systems Laboratory has initiated a program on fine particulate
control technology and some of the elements of the EPA program involve
some of the areas we identified as subjects for future research.

Recommendations for research programs in the areas of collection and
agglomeration mechanisms for particulates, conventional control systems,
and emerging and proposed control systems are discussed next.

COLLECTION AND AGGLOMERATION MECHANISMS

The review of collection and agglomeration mechanisms for particulates
served to pinpoint approaches and underlying mechanisms or forces which
should have potential for use in the control of fine particulates.
Table 1, page 2 lists the approaches that have significant potential.
The parametric dependence of many of the mechanisms or forces  underlying
the approaches listed in Table 1 are not well-defined and additional
research is warranted to develop requisite data.  Specific program areas
are discussed in the following subsections:

Studies of the Dynamics of Particulate Agglomeration

Small particles suspended in a gas can be subjected to a variety of forces
(e.g., thermal, electrical, sonic) which may cause them to come together
and form agglomerates.  Since particle agglomeration may'occur as a step
in a control system for fine particulates, a better understanding of the
dynamics of particle agglomeration under conditions suitable to industrial
gas cleaning applications is needed.  There are likely to be a number  of
factors of importance to the agglomeration process in specific cases.   In
the case of charged aerosols, electrostatic forces dominate and thus con-
trol the course of agglomeration.  For situations where steam is present,
condensation phenomena may control the process.
                                     8

-------
Identification of the dominant parameters in the agglomeration process
in specific cases of practical interest should be the objective of the
program.  Once the important parameters are delineated, methods for en-
hancing the agglomeration of fine particles should be defined and tested
in bench-scale equipment.

Sonic Agglomeration

Sonic fields are an effective way to agglomerate particulates.  Sonic
precipitators were tested extensively in the 1950's, and were shown to
effectively agglomerate very fine particulates.  These first-generation
devices used a sonic field to agglomerate fine particulates to a minimum
mean diameter of 10-15 microns.  The resulting particulates were then
removed from the gas stream by a conventional cyclone.  These early
sonic agglomerators had relatively large energy requirements—typically
0.2-0.4 hp/m3/min (5-10 hp/1,000 cfm).

The rate of agglomeration of particulates in a given sonic field is, to
a reasonable approximation, directly proportional to both the residence
time and the square of the particle density.  Consequently, if the re-
quired size of the final agglomerate could be reduced by an order of
magnitude, the energy requirements would be reduced by more than an order
of magnitude.  This reduction in required size could be achieved by the
use of an electrostatic precipitator to collect the agglomerates in place
of the cyclones previously used.  As a consequence, total energy require-
ments of 0.04-0.08 hp/m3/min (1-2 hp/1,000 cfm) or even less might be
attainable.

Work needs to be done to determine the energy required to agglomerate
various particulates to 1-2 micron mean diameter in sonic fields and the
resulting size distributions.  The efficiency with which electrostatic
precipitators can handle the agglomerated particulates also needs to be
determined since many "sonic agglomerates" have been reported to be
rather easily disintegrated.

Theoretical and Experimental Investigation of the Dynamics of Mixing
  of Steam. Water Droplets and Particulates

Steam and/or water drops can play an important role in enhancing several
collection mechanisms that might be exploited to improve the collection
of fine particulates (e.g., diffusiophoresis, steam condensation).  At
present, there is little more than an order of magnitude quantitative

-------
understanding of the phenomena occurring when steam, water droplets and
particulates are mixed.  This is especially true in equipment configura-
tions that are likely to be useful in actual industrial gas cleaning
applications.

The dynamics of the mixing process should be studied from both a theoreti-
cal and experimental standpoint.  Factors to be studied should include
the dynamics of mixing, nucleation phenomena, heat and mass transfer rates
and droplet and/or particle growth.  Configurations to be studied should
include cooled wall condensers, converging-diverging nozzles, sieve-plates,
jets and pipeline scrubbers.  The main thrust of the investigations
should be directed to the generation of basic information that will re-
sult in design equations for steam injection and mixing systems.

Particle Charging

The effectiveness of electrostatic forces for either particulate col-
lection or agglomeration is directly dependent upon the charge acquired
by the particulates.  Theory and experiment show that the failure of
particle charging systems to highly charge fine particles causes poor
fine particle collection in any electrical gas cleaning process.

Additional research on fine particle charging is recommended in order
to achieve the inherent potential of electrical gas cleaning processes.
Both theoretical and experimental research on fine particle charging
is recommended because present theories do not adequately describe fine
particle charging and existing experimental data are not representative
of actual gas cleaning systems.   Emphasis should be placed on the de-
velopment of charging systems that are practical for use with the various
types of electrostatically augmented control systems currently available
or under development.

CONVENTIONAL CONTROL EQUIPMENT

Conventional control equipment can play a significant  role in the con-
trol of fine particles—especially if the inherent  potential  of certain
types of devices can be realized.   Attainment of the full potential  of
conventional control equipment will require additional  research.   Recom-
mended areas of investigation are delineated below.
                                   10

-------
Electrostatic Precipitators

Theoretical studies and recent field testing have indicated that elec-
trostatic precipitators can be more effective for the removal of sub-
micron particulates than previously assumed.  Additional study of elec-
trostatic precipitators is recommended with effort focused on field
evaluations, to define their true potential for fine particulate collec-
tion on a variety of industrial sources.

Because the electrical characteristics of particulates emitted from many
sources fall outside the range where capture by dry electrostatic pre-
cipitators is most effective, research on wet electrostatic precipitators
is also recommended.  Initial effort on wet electrostatic precipitators
should involve a detailed feasibility study followed by field evalua-
tion tests on existing installations.

Fabric Filters

On those sources where they can be used, fabric filters are excellent con-
trol devices for fine particles.  Improvements in fabric filters which
would permit their use on a wider spectrum of emission sources would be
an important advance.  Recommended research topics are:  (1) feasibility
study of fabrics composed of finer fibers; (2) investigations of tech-
niques for tailoring of "active area" of filter fabric; and (3) high
temperature and high filtration velocity systems.

Wet Scrubbers

High energy wet scrubbers  (venturi) are capable of fine particulate
control on some sources.  Research on these devices should be directed
to:

1.  Optimization of condensation effects which may occur in venturi
scrubbers; and

2.  Development of new nozzle configurations which can both accelerate
and shear the water stream  into small drops at lower energy expenditures.
                                    11

-------
 EMERGING CONTROL SYSTEMS

 The  Steam-Hydro Air and ADTEC scrubber systems both offer promise as
 new  systems  for the control of fine particulate emissions in the near
 future.  Both of these systems appear applicable to a wide range of
 sources, and additional demonstration testing is recommended.

 PROPOSED CONTROL SYSTEMS

 Condensation scrubbers, charged droplet systems, electrified filters of
 various types, and foam scrubbers have sufficient potential for collecting
 fine particles to merit additional research and development.  Future
 studies on any of these systems should be conducted in equipment configura-
 tions and sizes that will produce reliable engineering design data.
 Recommendations for each system are outlined next.

 Condensation Scrubbers

 The  work on condensation scrubbing which has been conducted to date indi-
 cates it can be an important avenue to exploit in order  to collect fine
 particulates.  Previous work has shown that the specific details of heat
 and  mass transfer, the initiation of condensation, and various particle
 properties have very significant influence on condensation scrubbing.
 Current knowledge of the interaction of key parameters is minimal—espe-
 cially in equipment configurations that are likely to be useful in actual
 industrial gas cleaning applications.   Additional R&D is recommended to
 develop requisite data.

Multiple-stage or continuous-type equipment is  superior to  single-stage
equipment for both the  steam injection or sudden cooling approaches  to
condensation scrubbers, and multiple-stage equipment  is recommended  in
future studies.   In the steam injection mode, research should be  conducted
to determine:

 1.   Optimum steam requirements on each stage;

2.   Interstage cooling requirements; and

3.  Performance of multistage equipment with a  variety of particulates
representative of industrial emissions.
                                   12

-------
Research on sudden cooling should focus on determining the results of  the
addition of low-grade steam to hot gas streams followed by controlled
cooling.  Specific factors that should be defined are:

1.  Necessary thermodynamic conditions; and

2.  Optimum equipment configurations.

Charged Droplet Systems

Currently available data on charged droplet scrubbers and related devices
indicate some potential for the collection of fine particles.  However,
theoretical models generally predict poorer performance than has been ob-
served in the limited experimental work conducted to date.  Additional
work on the theoretical aspects of charged droplet scrubbing and on bench-
or pilot-scale equipment is recommended.   Emphasis should be placed on
experimental work on bench-scale or pilot-scale equipment so that mean-
ingful engineering design data are obtained.

Electrified Filters

Limited theoretical and experimental studies on idealized systems indi-
cate that electric fields do improve the collection efficiency of various
filter systems (e.g., fabric filters, wire mesh filters, granular beds,
fluid beds).  It is not clear whether electrified filters will be adaptable
to the control of emissions from industrial processes.

Research to explore the feasibility of this class of device  in more detail
is recommended.  The initial stage of the research should be a detailed
evaluation of the advantages and disadvantages of the various types of
electrified filters and the selection of the most viable for additional
evaluation.  Experimental evaluation under actual field conditions is
strongly recommended.

Foam Scrubbers

Hie addition of a surface-active agent to wet scrubbers has been reported
to increase the collection efficiency.  At present, the mechanisms of foam
scrubbing are ill-defined and  the utility of  foam scrubbing for  parti-
culate collection is difficult to determine.

Because the information on the mechanics  of foam scrubbing is meager,
research on all aspects of the process (i.e.,  foam generation, bubble
flow and dynamics, heat and mass transfer and  bubble  destruction)  is
recommended.

                                   13

-------
                            INTRODUCTION

As man has become more knowledgeable about the impact of particulate
pollutants, he has become increasingly aware of the importance to his
well being of the action of fine particles.  While our present under-
standing of the interaction between man and fine particulates  does not
permit a quantitative assessment of the overall influence of fine par-
ticulates , information continues to accumulate which points to many
negative features of fine particulate pollutants.

Control of the atmospheric loading of fine particulates poses a chal-
lenge because both primary fine particulates and secondary fine par-
ticulates (formed by complex transport and reaction processes in the
atmosphere) contribute to the overall atmospheric loading.  It is likely
that in order to insure the control of atmospheric fine particles which
exhibit adverse human health and welfare effects, control technology will
ultimately be needed to control both primary fine particulates and the
gaseous precursors of secondary fine particulates.

Midwest Research Institute (MRI), under a grant from the Air Pollution
Control Division, Office of Research and Development, Environmental
Protection Agency undertook a study to assess the state of the art of
control technology for fine particulates* emitted from stationary sources.
Emphasis was placed on the analysis of control technology for primary
particulates.

The initial activity in the program was directed to a general review of
collection and agglomeration mechanisms for particulates.  Both theoreti-
cal and experimental facets were reviewed with attention focused on
forces or collection mechanisms expected to be of importance for fine
particles.  The results of the review provided necessary background in-
formation and input for the subsequent analysis of control technology.
   There is no universally accepted definition of  fine particulates,
     and in this study, we have chosen an upper cut-off point of  3
     microns diameter.
                                    14

-------
The analysis of control technology for fine particulates was centered
on:  (1) conventional control equipment; (2) emerging control technology;
and (3) proposed or conceptual control systems.

The following sections of this report include:  (1) a discussion of the
highlights of our review of collection and agglomeration mechanisms
for particulates; (2) a delineation of promising mechanisms for fine
particulate collection or agglomeration; (3) the results of the assess-
ment of conventional, emerging and proposed control technology for fine
particulates; and (4) a series of appendices.
                                    15

-------
      COLLECTION AND AGGLOMERATION MECHANISMS FOR FARTICULATES

Farticulate collection is effected by passing the gas stream through a
system where the particles are acted on by forces which remove them from
the gas stream.  The forces acting  give the particles a component of
velocity in a direction other than that of the gas stream.  To be effective,
these forces must be sufficiently large to take the particle out of the
gas stream during its residence time in the system.  If the particulates
in the gas stream are submicron in size, their removal may be facilitated
by making agglomerates of the very small particulates and then collecting
the agglomerates.

The basic mechanisms or forces that can be used to collect or agglomerate
particulates are shown in Table 2.  The effectiveness of the individual
mechanisms is strongly dependent upon factors such as particle size,
flow velocity, particle density, particle number density, temperature
gradients, concentration gradients, and heat and mass transfer coefficients.
Individual collection or agglomeration mechanisms are briefly reviewed
next.

PARTICLE COLLECTION

Many of the forces or mechanisms in Table 2 are not effective in the
fine particle size range or do not produce a sufficiently large migra-
tion velocity to result in particle separation for realistic residence
times.  Appendix A presents an analysis of the effectiveness of the in-
dividual mechanisms or forces for the collection of fine particles.  The
main observations are presented in the following subsections.

Gravity and Momentum Forces

Although in theory a very large settling chamber would give sufficient
time for even small particles to be collected, practical size limita-
tions restrict the applicability of these chambers to large particles.
The efficiency of a simple settling chamber can be improved by giving
particles a downward momentum in addition to the gravity settling effect.
                                   16

-------
           Table 2.   PARTICLE COLLECTION OR AGGLOMERATION
                         FORCES (MECHANISMS)
I.  Particle collection

     A.  Gravity and momentum forces
     B.  Centrifugal forces
     C.  Aerodynamic capture

          1.  Inertial impaction
          2.  Interception
          3.  Diffusion
          4.  Electrostatic attraction
          5.  Gravitational settling

     D.  Flux forces

          1.  Electrostatic forces
          2.  Thermal forces
          3.  Diffusion forces
          4.  Magnetic forces

II.  Agglomeration and/or particle growth

     A.  Thermal or Brownian agglomeration
     B.  Turbulent agglomeration
     C.  Electrostatic agglomeration
     D.  Sonic agglomeration
     E.  Condensation
                                   17

-------
The number of possible designs incorporating this principle is very
great, varying from a simple baffle in the chamber to specially de-
signed jets which give accelerated settling.  While momentum separators
do improve collection efficiency, they do not significantly enhance
the collection of fine particles.

Centrifugal Forces

Particle removal from spinning gases by centrifugal forces is the method
by which cyclones collect particulates.  Theoretical and experimental
studies have shown that cyclone performance deteriorates rapidly as
particle size decreases below 5 microns.  The only use of cyclones in
regard to the control of fine particles might be as precleaners or in
connection with particle agglomeration devices.

Aerodynamic Capture

Aerodynamic capture of particles involves the collection of particles
by collecting bodies (e.g., droplets, fibers, packing, etc.).  To
utilize aerodynamic capture, the gas stream is brought near the col-
lecting bodies and then a number of short-range mechanisms accomplish the
actual collection.

In order to achieve a single particle target efficiency* of 90% or greater
for particles in the fine particle size range by inertial impaction,
velocity differentials in the range of 1Q2-104 m/sec (3.3 x 102-3.3 x 10*
ft/sec) and collecting bodies of 50 microns diameter or less are required
(see Appendix A).  To achieve the equivalent efficiency by interception,
the diameter of-the collecting body must be nearly the same as the par-
ticles to be collected.  For micron and submicron particles, this would
mean that the collecting body would have to be about 1 micron in size--
a virtual impossibility in practical situations.

Collection of small particles by Brownian diffusion is effective if suf-
ficient residence time is available.  If the collection of fine particles
by diffusion is to be a significant factor, the Peclet number (VgDc/DjjM)
cannot be much greater than 10.  In most practical situations, particle
collection by Brownian diffusion will be important only for particles
less than 0.5 micron.in diameter.
*  Single particle target efficiency is defined as

  _ cross-sectional area of fluid stream from which particles are removed
    cross-sectional area of filter element projected in direction of flow
                                     18

-------
Electrostatic forces between particles and collecting bodies can enhance
the collection of fine particles.  If the charge on the particles  is
relatively large, electrostatic forces can be many times greater than
mass forces.  Comparisons of the target efficiencies for electrical
forces with target efficiencies for inertial impaction and diffusion
show that for low gas velocities electrical collection may be much greater.
Electrostatic forces are likely to contribute significantly to aerodynamic
capture of particles in the 0.01-5 micron diameter size range when the
gas velocity is  low and the charge on the particles is relatively  large.

Flux Forces

Particles can be collected by  forces which result from electrical, tem-
perature, and concentration gradients, from the  flux of matter or  energy,
and from a magnetic field.  This group of forces, often termed flux
forces, are attractive for the collection of fine particles, because  the
magnitude of the flux forces do not approach zero as the size of  the  par-
ticles to be collected approaches  the submicron  range.  Theoretical
analysis indicates that, in general, the particle deposition velocity
resulting from flux forces should be of the order of 0.1 cm/sec (0.04
in/sec) or larger for apprecialbe collection efficiency in practical
equipment configurations.  Individual flux forces are discussed in the
following subsections.

Electrical Forces - If charged particles are subjected to an unidirectional
electric field,  they move towards the electrodes and are deposited.  In
the absence of turbulence or other aerodynamic effects, the migration
velocity  uu  resulting from the electrostatic force can be obtained from
Stokes' law and  is given by
                                  6nrpp


where  q  is the charge on the particle, E  is the strength of the elec-
trostatic field, rp  is the particle radius and  u  is the viscosity of
the gas.  This expression neglects second-order electrostatic effects
such as the polarizability of the particle  and assumes  that a spheri-
cal particle is moving in laminar flow (NRe < 1).  It is clear from Eq.
(1) that for a given particle size the only practical variables affecting
the migration velocity are the electric field strength and the charge
on  the particle.
                                    19

-------
Some fairly refined theories for particle charging mechanisms have been
developed, and these theories and available experimental evidence show a
minimum in the charging rate for particles in the 0.1-0.5 micron size
range.  However, the calculated particle-migration velocities for par-
ticles in this size range, obtained by assuming reasonable fields, charge
densities and residence times, are nevertheless quite large--of the order
of 1 cm/sec (0.4 in/sec).  The utilization of electrostatic forces to
collect fine particles is a promising avenue to pursue.

Thermal Forces (Thermophoresis) - Particles can be removed from a gas
stream by the use of a temperature gradient.  The force which causes par-
ticle motion results from momentum differences imparted to the particle
on opposite sides.  The hotter (and thus faster) molecules colliding with
the particle will impart a higher momentum to the particle than the cooler
(slower) molecules.  Aerosol particles will then drift in the thermal
gradient toward the cold surface.  The motion of aerosol particles as-
sociated with a temperature gradient is called thermophoresis.

A number of equations have been developed to calculate the velocity of a
particle due to the thermophoretic force.  All these equations can be
expressed in one general form
and differ in the definition of  K  .  The factor  K  allows for the ef-
fects of the thermal conductivities of the particle and gas and includes
various accommodation coefficients which are not readily measured or
evaluated a priori.  In simple theories  K  is independent of particle
size, while in the more complicated theories  K  also depends on particle
size.

The thermophoretic velocity is quite low, being of the order of 0.05-
0.1 cm/sec (0.02-0.04 in/sec) even at temperature gradients of 103°C/cm.
Long residence times would be required to achieve significant particle
collection at more reasonable temperature gradients.  The large space
requirements coupled with the high cost of maintaining the required
temperature gradient severely restricts applications of thermal pre-
cipitators.  However, thermal forces might be exploited to enhance the
performance of certain types of control devices such as packed or granular
beds.
                                    20

-------
Diffusion Forces  (Diffusiophoresis)  -  In a  concentration  gradient,  which
is  accompanied by diffusion but not  necessarily by net motion of  the gas
phase,  the heavier molecules will  impart a  higher momentum than the
lighter molecules.   If  there is a  net  motion  of the  gas phase (Stefan
flow),  additional force is applied to  the particles.  The combination of
forces  due to Stefan flow and  the  concentration gradient  is referred to
as  the  diffusiophoretic force.  Particle movement by this force is  called
diffusiophoresis.

To  exploit the mechanism of diffusiophoresis,  the collecting device must
be  designed such  that one component  in the  gas phase is diffusing toward
a collecting surface.  The most practical case is  the diffusion of water
vapor toward a surface. References  1, 2, and 3 indicate  that for water
vapor diffusing  through air at 0°C and 1 atm  total pressure, the  particle
diffusiophoric velocity is given by
                         vp  =  -  1.9  x  ID"* g                   (3)


 Theoretical  analysis recently completed by Calvert, et al.,4/ indicates
 that  positive diffusiophoresis  (i.e., condensation) can produce particle
 deposition velocities of 0.1  cm/sec (0.04 in/sec) or larger under the
 heat  and mass transfer conditions of a realistic scrubber.
      i

 Magnetic Forces - A force (Lorentz  force) is generated when an electri-
 cal charge or an electrically charged particle moves in a magnetic field
 transverse to the field  lines.  If  a dust particle carrying "n" elementary
 charges "q" moves with a speed  v , the direction of the force will be
 at right angles to both  the direction of the field and the direction of
 motion of the particle so that  the  particle will be diverted from its
 original path.  As a result of  the  change in direction of the particle,
 the possibility of particle precipitation exists.

 Equation 4 presents an expression for the terminal drift velocity of a
 charged particle in a magnetic  field
                             vm - S*     .                  (4)
Equation 4 indicates that the terminal drift velocity varies directly
with the number of charges on a particle, the particle velocity, and the
                                    21

-------
magnetic  field  strength, and  inversely with  the  particle diameter.  Ex-
amination of Eq.  (4)  indicates  that  the  terminal drift velocity  is  less
than  0.01 cm/sec  (0.004 in/sec) for  particles 0.1-1.0 microns  in size
except at very  high particle  velocities  and magnetic field strengths.
The above theoretical considerations demonstrate that for all  practical
purposes  the force generated  by a charged particle moving in a magnetic
field is  too small for use in particle separation for industrial gas
cleaning.

PARTICLE  AGGLOMERATION AND GROWTH

The agglomeration or  growth of fine particulates to particles  of 2-5
microns in size, as a step in the control of fine particulates,  is  in
principle attractive.  If large particles can be produced for  reasonable
energy expenditures,  it would be possible to utilize devices which  cause
particle  agglomeration or growth in conjunction with conventional control
systems.  Several possibilities exist for particle agglomeration, and
condensation of water vapor on particles is the most feasible way to
achieve particle growth without agglomeration.   The utility of agglomera-
tion  and  condensation processes are discussed in the following subsections.

Thermal Agglomeration

When  the movement of  particles  leading to contact and agglomeration
can be accomplished only by Brownian diffusion,  the process is called
thermal agglomeration.  Thermal agglomeration has a very serious dis-
advantage with regard to particulate collection—the long time required
to achieve significant particle growth.  The only variables which can be
used  to control particle-particle collision rate are gas temperature,
gas composition, and  the particle concentration.  Gas temperature and
gas composition are not very useful variables.

One could conceivably change the particle distribution and concentration
by seeding the particle suspension with large particles to act as ag-
glomeration sites, but this is not very useful because the total agglom-
eration rate depends  inversely on the square of  the particle number
density.  Thus, thermal agglomeration is not a viable approach to par-
ticle growth for the  purposes of control of fine particulates.

Turbulent Agglomeration

Turbulence increases  the relative velocities among particulates which in
turn  increases the chance of particle collision.  Theoretical studies and
limited experimental  investigation have shown that the rate of dissipation
                                    22

-------
of  turbulent energy per gram of medium must be quite high for the agglom-
eration rate of micron and submicron particles to be accelerated noticeably
above  thermal agglomeration rates.

Theoretical studies and available experimental data indicate  that turbulence
does not  significantly enhance the agglomeration constants  for  particles
less than 0.2 micron  in diameter, but can increase the agglomeration  con-
stant  of  0.5 micron particles by about a factor of 10 and 1 micron  par-
ticles by a factor of 102.  The power expended to accomplish  this increase

would  be  of the order of  107 dvne-cm-sec    ^ 40 hp/1,000 ft3).  This is
                                 gram
a very high energy consumption rate and, therefore, turbulent agglomera-
tion does not appear  attractive if an external source is needed to  pro-
vide the  turbulence.  If  a stream is naturally highly turbulent, attempts
to  foster  turbulent agglomeration might have  some merit depending upon
the control system used to collect particles  after the agglomeration  step.

Sonic  Agglomeration

A potentially effective way of agglomerating  fine particles is to pass
the aerosol through a sonic field.  Several different effects are respon-
sible  for enhanced particle agglomeration due to sonic forces, including:
(1) collection of particles at the antinodes  in the sonic standing wave
(radiation pressure); (2) hydrodynamic forces between the particles;  and
(3) additional collisions due to the different vibrational amplitudes of
different sized particles.  Although no comprehensive theory is available
for sonic agglomeration,  some idea of the parametric dependence can be
discerned from equations which depict each separate mechanism (see Appendix
A).

All three forces show a strong dependence on  the radii of the particles,
with the force due to radiation pressure and hydrodynamics increasing
with particle size and the relative vibrational amplitudes decreasing
with increasing particle  size.  There is also a strong frequency de-
pendence for all three forces.  At low frequencies all forces are negli-
gible,while at high frequencies the vibrational amplitude of the particles
become negligible.  There is, therefore, an optimum frequency for acoustic
agglomeration which varies with particle size.

There  is a rather large body of diverse experimental data on sonic ag-
glomeration.  Reference 5 presents an extensive review of both experi-
mental data and industrial experience.  Several generalizations can be
made from this body of literature.
                                    23

-------
1.  Sonic fields can effectively induce agglomeration in all sizes of
suspended particles, including those in the submicron range.

2.  All materials can be agglomerated.

3.  The efficiency of the agglomeration depends upon the square of the
particulate concentration.  The minimum particulate loading, that can
be treated, is usually given as 1.1-2.3 g/m3 (0.5-1.0 grains/ft3), al-
though number density and mean size would appear to be more important
parameters.  For more disperse particulate suspensions it is necessary
to seed the system or to add a mist to bring the particle loading to a
satisfactory level.  The addition of mist is quite inefficient because
some of the mist added will also agglomerate and disappear from the sys-
tem before particle agglomeration is achieved.

4.  The sonic agglomeration process is strongly frequency dependent, with
the optimum frequency varying with the particle size or sizes to be ag-
glomerated.  Optimum frequencies range from a few hundred to an upper
limit of about 10,000 Hz.

5.  The required residence time for a given degree of agglomeration varies
with field intensity, and minimum field intensities of about 140 dB are
required for usable residence times.  For a 150-160 dB field a residence
time of 5 sec is typical.

Agglomeration of Charged Particles

One method of increasing the rate of agglomeration of fine particulates
is to add a bipolar charge, either with or without an externally imposed
field.  If relatively large electrostatic forces between particulates can
be produced, a significant increase in the rate of agglomeration of sub-
micron particulates can result.

Fuchs-' and Zebel—' have shown that the influence of particle charge on
agglomeration rates can be expressed in terms of a correction factor,
 kem  to the agglomeration constant in the absence of particle charge,
 KQ .  The correction factor, kem  > is a function of the charge on in-
dividual particles and the particle size of individual particles.  For
weakly charged monodisperse aerosols (see Figure A-14, Appendix A), there
will be little change in the agglomeration rate for fine particles.  For
highly charged aerosols, the agglomeration rates can be much higher (up
to 10^-10^ times that of the uncharged aerosols), but will usually fall
rapidly because of annihilation of the charge upon agglomeration and the
decrease in number density.  This result could be unfavorable because it
may not be possible to achieve significant growth of fine particles.

                                    24

-------
However, if we consider a bipolarly charged polydisperse aerosol con-
taining a significant number of relatively large particulates as well
as the fine particulates, the situation appears much more favorable be-
cause it may be possible to collect the fine particulates on the large
particulates.  If the particle charge on an individual particle is pro-
portional to the particle surface area and if the small particles cover
the large particle as a coating no more than one layer thick, the number
of small particles which can be deposited on the large particle before
charge neutralization is of the order of

                                                .2
                     diameter of large particle \
                     diameter of small particle/

Thus, the neutralization of a large particle will require a substantial
number of small particles and the fine particle concentration in a gas
could be significantly reduced before charge annihilation.  The utiliza-
tion of a bipolarly charged polydisperse aerosol to effect particle ag-
glomeration is a promising avenue for additional study.

Particle Growth by Condensation

The size of particles can be increased by causing a vapor to condense on
the surface.  The most practical vapor to utilize for this purpose is
water.  Nucleation and condensation processes for water have been studied
in some detail.  Water vapor may self-nucleate and condense into drops in
the absence of condensation nuclei if the degree of supersaturation is
great enough.  At lower supersaturations, however, a nucleating surface
is necessary for condensation.  The conditions necessary for condensation
to occur on a surface are determined by applying equilibrium thermodynamics,
If an insoluble surface is present, drops of the critical radius will form
on this surface at a higher rate than they will self nucleate.  This is
due to the fact that they have less volume and require fewer molecules
than for the case of homogeneous nucleation; hence the free energy in-
crease is smaller.  Their volume is smaller because they can form as a
partial sphere resting on the surface rather than the complete sphere
necessary in the absence of the surface.   If the particle is  soluble,
nucleation will occur even more readily

After a drop of critical size has been nucleated, it grows  at a rate
determined by the ambient environment and the conditions  at its sur-
face, just as the condensation flux to a flat wall depends  on these
                                    25

-------
boundary conditions.  Various equations have been developed to describe
the rate of growth and, in general, they all require knowledge of the
rate of mass transfer, rate of heat transfer, and phase equilibrium.

If particle growth of fine particles due to condensation can be achieved
and particles grown to 2-5 microns, subsequent collection of the drop-
lets could be achieved in conventional equipment.  Thus, particle growth
has the advantage that high particle deposition rates can persist after
heat and mass fluxes are dissipated because once the particles are en-
larged they can be collected at little additional pressure drop.

SUMMARY

The review of collection and agglomeration mechanisms for particulates
pinpointed mechanisms or forces which should be the most useful for the
control of fine particulates.  Foremost in this regard are inertial im-
pact ion, flux forces, electrostatic agglomeration and sonic agglomera-
tion.  Some of these mechanisms or forces are already employed in con-
ventional control equipment, but in general in a somewhat inefficient
manner.  In some instances it may be possible to modify or redesign con-
ventional control systems to enhance the contribution of the promising
mechanisms and improve fine particulate removal.  Alternatively, new con-
trol systems will have to be developed.  The following sections of this
report discuss:  the current capability of conventional control equip-
ment to collect fine particles and avenues to pursue to improve per-
formance of conventional equipment,and the capability of recently de-
veloped or proposed systems for the collection of fine particles.
                                    26

-------
     FINE PARTICLE COLLECTION BY CONVENTIONAL CONTROL EQUIPMENT

The status of conventional control equipment was reviewed in order to de-
fine the extent of our existing capability in regard to the control of
fine particulate emissions.  A secondary objective was to identify pos-
sible avenues for improving the performance capability of conventional
equipment with emphasis on the collection of fine particles.

While conventional control devices do in fact collect fine particulates
in some applications, the performance of conventional control equipment
is often not satisfactory.  The poor performance with regard to fine
particulates results from:

1.  Inherent limitations of conventional control equipment in the fine
particle size range, and

2.  Lack of previous need to control fine particulates and the attendant
lack of interest in optimizing systems with emphasis on the collection of
fine particles.

Item 1 results from the fact that the mechanisms or forces exploited in
a specific device are not effective in the fine particle size range.
The second item, which is only applicable to control systems not in-
herently limited by the forces or collection mechanisms utilized in the
systems, is a result of previous priorities established for the collec-
tion of particulate emissions.

Exceptions to the preceding comments exist where conventional collection
systems have reached very high mass efficiencies resulting in significant
capture of fine particulates along with large particulates.  Installations
of high efficiency electrostatic precipitators, fabric filters, and high
energy wet scrubbers have been reported which achieve relatively high
collection of fine particulates.11.33-34/  it would appear that at least
some combinations of conventional control systems and source characteris-
tics can result in significant control of fine particulate emissions.
                                    27

-------
The analysis of conventional control equipment performed in this study
was focused on systems deemed capable of significant collection of fine
particulates, i.e., electrostatic precipitators, fabric filters, and
high efficiency wet scrubbers.  The principal findings of the review are
highlighted in the following sections.

ELECTROSTATIC PRECIPITATORS

Electrostatic precipitators utilize the migration of charged particles
in an electric field for the collection of particulates.  In the absence
of turbulence or other aerodynamic effects, the migration velocity  &
resulting from the electrostatic force is given by Eq. (1), p. 19.

The potential usefulness of a "conventional" electrostatic precipitator
for the control of submicron particulates can be send-quantitatively
assessed in terms of the migration velocities and the Deutsch equation^'


                            f = i_e-*»/v                       (5)
where  f  is the fractional efficiency for a particle with a migration
velocity  u>  of a precipitator with a collecting surface with an area  A
and a gas volume flow rate  V .  Although the Deutsch equation is not
rigorous, it is acceptable for the purposes of this discussion.

Modern high efficiency electrostatic precipitators are often designed
with 1.3-2.0 m2 of collection area per cubic meter per minute of gas
flow (400-600 ft2/!,000 cfm),* and operate with fields in excess of
10^ v/m and charge densities of 2-5 x lO1^ ions/m^.  Particle migration
velocities of the order of 2 cm/sec for 0.2 micron diameter particles
and 3 cm/sec (1.2 in/sec) for 1 micron diameter particles can be esti-
mated under these conditions (see Figure A-6, Appendix A).  Fractional
efficiences calculated from Eq. (5) using these migration velocities and
A/V  ratios are shown in Table 3.
   Electrostatic precipitators currently designed with 400-600 ft2/l,000
     cfm are typically for flyashes having high resistivities.
                                    28

-------
           Table 3.  ESTIMATED FRACTIONAL EFFICIENCIES OF
                     ELECTROSTATIC PRECIPITATOR^'
       A/V
(min/m)    (min/ft)
  1.48
  1.64
  1.80
  1.97
0.45
0.50
0.55
0.60
            Up
          (micron)
0.2
                               U)
(cm/sec)    (in/sec)

   2          0.79
Fractional Effi-
  ciency (%)

       83
       86
       89
       90
  1.48
  1.64
  1.80
  1.97
0.45
0.50
0.55
0.60
1.0
              1.18
       93
       95
       96
       97
a/  Electric field - 2 x 1Q5 v/m.
    Charge density - 2-5 x 1013 ions/m3.
                                   29

-------
The general validity of these very crude calculations is suppored by
recent field testing of high-efficiency electrostatic precipitators with
advanced particle sizing techniques.  The data shown in Figure 1, ob-
tained by Southern Research Institute, illustrate this point.  Figure 1
presents measured and computed efficiency as a function of particle size
for a 99.6% overall efficiency ESP operating on a coal-fired power plant.
Cascade impactors and a diffusion battery were used to obtain the frac-
tional efficiency data.  Collection efficiencies in excess of 90% on a
mass basis were noted for particles in the 0.1-1.0 urn size range.  On a
number basis, the electrostatic precipitator is removing 95% of the
par tides ..2'

Similar fractional efficiency characteristics for electrostatic precipi-
tators have been reported by European investigators.  Figure 2 presents
some data reported in Ref. 10.  Neither the type of source being con-
trolled nor the sampling procedures used to obtain the fractional ef-
ficiency data were reported in Ref. 10.

Unfortunately, the electrical characteristics of particulates emitted
from many specific sources fall outside the range where capture by elec-
trostatic precipitation is most effective.  Modifications to the designs
of electrostatic precipitators or conditioning of the particulate emis-
sions (or both) could be necessary to capture fine particulates in those
cases.

The important point which emerges from the preceding calculations and
comparison with recent field testing is that combined increases in the
size of electrostatic precipitators, the charge density, and the field
strength can produce collection efficiencies of 90% or greater for fine
particulates.  These changes are within the present state of the art and
should be economically feasible.  Furthermore, it appears possible that
equivalent effects can be obtained by using precipitators with multiple
stages and/or increased sectionalization.  Conditioning of the particulates
might also be used to enhance collection efficiency.

Consideration should also be given to the further development of wet
electrostatic precipitators.  Exit grain loadings as low as 0.002
 (°
0.001 Srains )  and overall collection efficiencies of 99.9% have been
       ft3  /
reported for wet electrostatic precipitators.££/  Although not verified
by careful  testing, high efficiency for submlcron particulates is also
claimed at  low power consumption.  Water usage rates are reported to
                                    30

-------
 c
.2
'o
a:
99.9
99 8
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5
0.2
0.1

0.01
0.





^V^
• ^v^^^
* " 	













	 L 1 I 1 1 I , .
01 o.
	 1 	 1 	 1 — 1 — 1 1 1 1


m
A A A
• • •
	 *
- 	 -r^
f
fYimputerl













1 1.
• 1 1 	 1 1 1 1 I1

. Y
•* /
A * /
" A
^r Computed
S














0 1


















Particle Sizing
Method

Oitfusional
A Inertia 1
• Optical
D
                                            Diameter, [4.m


 Figure  1.  Measured and computed efficiency as a function of particle size  for electrostatic precipitator

                                     installation at a coal-fired power  plantE/

-------
IUU
0)
u
c£
X
u
S 75
[o
UJ
50
0.
-ft -©""O^™""^"""^
"" ^N/
H

till
01 0.1 1.0 10 10
                           Particle Diameter, Microns
Figure 2.  Variation of electrostatic  precipitator efficiency with size
             of particle dust  (sizing  by electron microscope)—
                                                                '
                                    32

-------
range from 0.07-1.3 liters/m3/min (0.5-10 gal/1,000 cfm) off recirculating
water, depending upon equipment configurations.  Pressure drops ranging
from 0.2-2.5 cm H20 in the electrostatic section have been reported.

One of the most important advantages of wet systems in comparison to dry
units is the efficient collection of a wide range of dusts.  Since the
particles are generally thoroughly wetted in this type of precipitator,
they are preconditioned and are effectively charged and precipitated,
regardless of the degree of resistivity of the material.  Particle re-
entrainment would also be essentially zero in these systems.  Wet elec-
trostatic precipitators also generally operate at gas velocities of
higher magnitude than those in conventional dry precipitators. For this
reason, the device is much smaller, more compact, and requires less in-
stallation space than does conventional equipment, thereby reducing the
capital cost.  The major disadvantages of wet electrostatic precipitators
are the need to handle a wastewater stream and a possible increase in
equipment corrosion.

FABRIC FILTERS

On those sources where they can be used, fabric filters are excellent
control devices for fine particles.  Our review of the theoretical models
of fabric filtration led to the conclusion that existing models are of
limited use in assessing possible avenues for improving fine particle
collection by fabric filters.  The relatively limited utility of existing
models arises from the fact that the models are usually based on single
fibers or incorporate the assumption that the filters are sufficiently
clean for no deposition to take place on previously deposited particles.
The latter assumption is the most serious limitation with regard to in-
dustrial fabric filter systems.

A layer of deposited particles, called the filter cake, forms part of the
filtering media in an industrial fabric filter.  While the deposit is
building up and the filter is becoming clogged, through use, important
changes may take place which considerably alter its filtering characteris-
tics.  There appear to be two main factors.  First, the distribution of
pore size may be wide so that flow takes place preferentially through
coarse pores and the filtration efficiency is low.  When the coarse pores
become clogged, to a certain extent, the pore distribution becomes more
uniform and the median pore size is smaller so that both flow resistance
and collection efficiency rise quickly.
                                    33

-------
Secondly,  the deposited particles do not as a rule distribute  themselves
evenly over the surface of  the fibers, but build up chain aggregates which
act themselves as very fine fibers and may collect particles more ef-
fectively  than the material of which the filter is made.  Available in-
formation  indicates that particles which choke a filter most rapidly have
acicular shape, and that small particles act more effectively  than large
ones.

Our review has demonstrated clearly that the performance of fabric filter
systems are dependent upon  the interaction between particle properties
and the engineering parameters of the filter system.  The important
properties of the particulate matter include the diameter, size distri-
bution, shape, surface characteristics, electrostatics, chemical reactiv-
ity, adhesiveness and hydroscopicity of the particles.  They affect the
structure  of the deposited cake; its flow resistance; the power required
for filtration; the rates of plugging and blinding; and the rate of mechani-
cal abrasion of the fabric.  In addition, they influence the particulate
penetration through the filter, and thus, affect the collection efficiency
of the systems.

The air-to-cloth ratio, cleaning mechanism, temperature, humidity, weave
pattern, fabric weight, fiber diameter, gas flow rate and filter fabric
"surface" characteristics appear to be the most important engineering
parameters.  The filter fabric "surface" is defined as that region of the
fabric which,  in successive loading and cleaning cycles, has a significant
influence on deposition and removal characteristics.   Little appears to
be known about the electrical, geometric, adhesive and mechanical properties
of the filter fabric "surface" despite the fact that the interface between
the dust and fabric "surface" is, perhaps, the most important facet of
the entire fabric filter system.  A conclusion of the analysis of fabric
filtration is  that improved collection of fine particulates might well
result from a better understanding of the interface between particulate
properties and the characteristics of the filter fabric "surface."

A potential negative feature that may result from improved collection
of fine particles would be an increased difficulty of cleaning fabric
filters with an attendant increase in pressure drop during normal operating
cycles.  Currently, fabric filters on industrial installations typically
operate with a pressure drop of about 10.2 cm wg.  (4 in. wg.).  The
maximum pressure drop for effective operation has  not been defined,  and
with modifications to fabric filters high pressure drop systems may be
feasible.
                                    34

-------
Fabric filter systems currently available have inherent limitations which
preclude their use on a variety of sources of particulate pollutants.
In general, fabric filters require a rather low linear gas velocity for
effective operation.  Also, most filter media must be used at relatively
low temperatures and on a relatively dry gas in order to function properly
and avoid excessive deterioration.  Many also have rather limited re-
sistance to corrosion and rather low mechanical strength, which limits
the amount of handling they can withstand.  Research and development work
to extend the range of source applications of fabric filters should be
given consideration, especially the development of high temperature and
high filtration velocity systems.

WET SCRUBBERS

The collection of particulates by wet scrubbers results from a variety
of interactions between the gas and liquid phases.  Wet scrubbers  cover
a wide range of designs, sizes, and performance capability.   No single
mathematical model has been developed which will adequately predict the
operating characteristics or efficiency for this broad class of collectors.
In the past, even for a single type of scrubber the design equations have
not been too accurate and are generally empirical in nature.

Recently Calvert, et al.,—' have used the following categories to de-
scribe the various types of scrubbers:  plate, massive packing, fibrous
packing, preformed spray, gas atomized spray, centrifugal, baffle, im-
pingement and entrainment, mechanically aided moving bed, and combination.
Calvert has used an approach to understanding and designing scrubbers—
the unit mechanism—that is extremely useful in trying to predict ways
in which scrubber efficiency can be improved for collection of micron and
submicron particulates.

Examples of unit mechanisms include:

Scrubber Type               Unit Mechanisms

Venturi                Liquid jets, sheets, and drops

Sieve plate            Gas jets and bubbles

Impingement plate      Gas jets, jet impingement, liquid drops
                         and sheets

Mobile packing         Gas jets, liquid jets, drops and sheets

Packed column          Liquid sheets and drops, curved liquid
                         sheets, gas jets, and bubbles
                                    35

-------
To determine the scrubber collection efficiency using the unit mechanism
approach' requires consideration of transport phenomena such as diffusion
(molecular and turbulent), inertial impaction, interception, electro-
static attraction, diffusiophoresis, thermophoresis, etc.  The effectiveness
of a given scrubber in collecting submicron particulates can be estimated
from the unit mechanisms and the transport phenomena involved.

High energy wet scrubbers (i.e., venturi and disintegrator types) are
capable of fine particulate control.  The high energy scrubbers utilize
inertial impaction as the principal collection mechanism.  The effective-
ness of inertial impaction as a collection mechanism decreases markedly
at particle sizes below 1 micron unless high velocity differentials are
maintained between the collecting bodies and the particles.  The con-
tribution of inertial impaction could also be enhanced by decreasing the
size of the collecting bodies.  In order to maintain or increase velocity
differentials and decrease the size of the collecting bodies (water drops),
energy consumption will have to be increased.  Thus, while there may be
some possibility for improving fine particle collection by conventional
venturi type scrubbers, even higher energy consumption is likely.

Improvement in collection efficiency for fine particles of other types
of conventional wet scrubbers will come from designs that take better
advantage of flux forces--thermophoresis, diffusiophoresis, etc.

SUMMARY

Well-designed conventional control systems (i.e., systems optimized with
respect to the collection of fine particulates) are capable of achieving
relatively high collection of fine particulates.

High-efficiency electrostatic precipitators appear to offer considerable
potential in this regard.  Changes within the present state of the art
(i.e., increases in the A/V ratio, charge density, and field strength)
could produce significant improvement in the collection of fine par-
ticulates by electrostatic precipitators.

Improvements in fabric  filter technology might also result in better
control of fine particles from some sources.  Avenues to pursue to im-
prove performance are not as clear-cut as with electrostatic precipitators.
Improvements in fabric  filters which would permit their use on a wider
spectrum of emission sources would be an important advance.
                                    36

-------
New designs for high-energy wet scrubbers which produce smaller water
droplets and higher differential velocities at energy expenditures
equivalent to or lower than currently available systems would be an
avenue to pursue.
                                   37

-------
                EMERGING AND PROPOSED CONTROL SYSTEMS
                        FOR FINE PARTICULATES

A variety of devices, touted as being effective for the collection of
fine particulates, have been reported in the literature.  The devices
range from concepts or proposals to laboratory-scale devices to full-
scale units that have been tested on industrial sources of fine partic-
ulates.  During this study attempts were made to obtain all readily
available information on many of these devices in order to evaluate their
potential.  The technical staff of the Particulate Technology Section,
Particulate and Chemical Processes Branch, Control Systems Laboratory
(CSL), provided valuable technical information and other data on several
of the systems.  Research, field testing, and theoretical evaluations
have been or are currently being conducted under CSL sponsorship on some
devices discussed in this section.

Appendix B discusses some of the emerging systems, while Appendices C-I
present evaluations of concepts or proposed systems.  Table 4 delineates
the systems reviewed in these Appendices and the highlights of the re-
view are discussed next.

EMERGING CONTROL SYSTEMS

The emerging control systems listed in Table 4 are those systems for
which sufficient data could be obtained to make an assessment of perfor-
mance capability.  The individual systems are discussed in some detail in
Appendix B.  Of the systems analyzed, the Steam-Hydro Air Cleaning and
the ADTEC Wet Scrubber were judged to have the most immediate potential
for use as fine particle control systems.  Wetted-knit mesh separators
also appear to have potential on some applications.  These systems are
discussed briefly in the following sections.
                                   38

-------
           Table 4.   EMERGING AND PROPOSED  CONTROL SYSTEMS
                        FOR FINE PARTICULATES
I.  Emerging Control Systems

     A.  Steam-Hydro Air Cleaning
     B.  ADTEC Wet Scrubber
     C.  Cross-Flow Nucleation Scrubber
     D.  PENTAPURETM IMPINGER
     E.  Dynactor Wet Scrubber
     F.  Mystaire™ Scrubber
     G.  Dupont Wet Scrubber

II.  Proposed Control Systems

     A.  Granular bed filters
     B.  Fluidized beds
     C.  Condensation scrubbers
     D.  Foam scrubbers
     E.  Charged droplet scrubbers
     F.  Electrified filters
     G.  Gamma-ray precipitator
     H.  Sonic agglomerators
                                  39

-------
Steam-Hydro Air Cleaning System

Lone Star Steel has  recently  developed  a wet-scrubbing system which uti-
lizes a high-speed steam drive with injected water to  accomplish the re-
moval of particulate pollutants.  The system consists of a steam nozzle,
water injector, mixing tube, and twin cyclones.  (See Figure B-l,  Appendix  B.)
Normally, the system operates on energy produced by waste heat captured
from the process being controlled.  The heat is used to generate steam
in a waste heat boiler.  In installations where little or no waste heat
is available, supplemental heat may be provided.   In many cases, a
package steam boiler may supply all  the energy.

The steam nozzle, a proprietary design, surrounded by eight spray nozzles
(see Figure B-2, Appendix B)  is a key element in this system.  The exit
velocity of the steam is about 1,050 m/sec  (3,500  ft/sec).  The function
of the steam is to break up or shear the surrounding water spray and to
provide for moving the gas stream through the system.  The flow rate of
injected water, controls the cleaning efficiency.  The steam flow rate has
little to do with the cleaning efficiency, but some is required to shear
the injected water spray.  Particle collection occurs primarily by
inertial impaction.  Since the velocity of the collecting bodies is
about 10^ m/sec, single target efficiency of 90% or greater would be ex-
pected (see Figure A-2, Appendix A).

Steam requirements, as reported by Lone Star Steel, range from 0.03-0.16
kg of steam per cubic meter of gas (2-10 lb/1,000 ft3), depending primarily
on upstream process draft requirements.  Water rates are 0.23 kg of H20
per kilogram of gas (230 Ib of H20/l,000 Ib of gas) for the conditioning
or atomizing chamber and for  the spray nozzles surrounding the steam jet.

Efficiency tests have recently been performed by a contractor* funded by
CSL on units installed on multiple open-hearth (oxygen lanced) steel
furnaces at Lone Star Steel.  Preliminary test results indicate collec-
tion efficiencies of 99.9% or greater.  During the tests, inlet grain
loadings averaged 2.88 g/m3 (1.26 grains/scf), but varied from 0.23-6.86
g/m3 (0.1-3.0 grains/scf) depending on the point in the process cycle.
Outlet grain loadings ranged from 0.002-0.02 g/m3  (0.0008-0.0089 grains/
scf).  The inlet size distribution of particles had a mass mean diameter
of 1.5 microns and nearly all the particulate was less than 3 microns in
diameter.!/
   Tests performed by Southern Research Institute using cascade impactors
     and a diffusion battery for, particle sizing and fractional effi-
     ciency measurements.
                                   40

-------
Capital, operating, and maintenance cost data are not available for the
device.  Lone Star Steel has stated that "the installed cost is more than
competitive with conventional devices (excluding cost of waste heat
boilers)."^

Based on the results of the preliminary tests, the Steam-Hydro system
offers promise as a new system for the control of fine particulate emis-
sions.  The system appears to be applicable to a wide range of sources,
but primary applications are likely to be sources where waste heat is
available and which emit fine participates.  Where waste heat is not
sufficient to power the system, supplemental heat may be used.  However,
the use of supplemental heat would add to the overall cost of the system.

ADTEC Wet Scrubber

The ADTEC system, developed by Aronetics, Inc., Tullahoma, Tennessee,
is a wet scrubbing system that operates on the conventional venturi col-
lection-mechanism of inertial impaction, but establishes the requisite
particle-droplet differential velocity by utilizing waste process heat
rather than external energy.11'  The Aronetics system is shown schemati-
cally in Figure B-4, Appendix B,  Waste gas from the industrial process
passes through the heat exchanger, the two-phase jet, and the separator,
in that order.  Water is pumped at 300 psi to a heat exchanger where
heat is transferred to the water from the gas, increasing the temperature
of the water to the required level, usually between 300°F and 400°F.
Expanding through the nozzle of the two-phase jet, partial flashing occurs
(approximately 15% vaporizes)  and the remaining liquid is atomized
(water droplets average approximately 10 microns).  Thus, a two-phase
mixture of steam and small water droplets leaves the nozzle at high
velocity.  The dirty gas is entrained by this high velocity two-phase
mixture, and in the ensuing mixing with the gas, cleaning occurs primarily
by impaction.  At the same time, transfer of momentum to the gas results
in a pressure rise across the mixing section.  This pressure rise pro-
vides the motive force to move the gas through the system.

The mixture of steam, gas, and entrained water with dust collected in
cleaning  passes through the separator after discharging from the mixing
section.  The water with the dust is removed from the gas and steam mix-
ture which then leaves the separator through the stack.  The water with
dust is discharged from the separator to the wastewater treatment sub-
system.  Chemicals and other treatment are applied to remove the dust and
other contaminants from the water for recovery of the water and to prevent
scale and corrosion in the system.
                                   41

-------
The treated water is returned to the pump for recycling and sludge is re-
moved for disposal by other pumps not shown.  Make-up water to replace
water leaving the system through the stack in vapor form and with the
sludge is introduced in the separator where some heat is transferred to
the make-up water from the gas.

Since the transfer of momentum produces a pressure rise across the mix-
ing section which provides the motive force to move the gas through the
system, the ADTEC system does not require, nor does it include, a blower
which is necessary in other cleaning systems or methods.

A test unit was installed on a ferro-alloy furnace of the Chromium Mining
and Smelting Corporation (Chromasco) at Woodstock, Tennessee.  Test re-
sults from the pilot system are given in Table B-l, Appendix B.  Collec-
tion efficiencies of 99+% were reported.  Since the device is similar
in some respects to the Lone Star Steel system, collection efficiencies
of the two systems should be comparable.  The Control Systems Laboratory
is currently planning to fund additional testing of the ADTEC system using
the latest techniques for particle sizing and fractional efficiency
measurement.

Mechanical energy consumed in pumping the water to the required pressure
of 275-300 psi is about 0.1 hp/m3/min of gas (3 hp/1,000 cfm), while the
power required in the form of heat input to the water from the gas stream
exceeds 7 hp/m^/min (200 hp/1,000 cfm).  The latter power requirement
would obviously be excessive if it were not supplied by waste heat from
the gas stream.  Capital costs have been estimated by Aronetics to be
between those of venturi scrubbers and fabric filters.

On the basis of currently available information, this system also appears
to offer significant improvement in the collection of fine particles at
modest energy consumption rates  where a waste gas is available which
contains a sufficient amount of thermal energy to heat  the water.

Wetted-Knit Mesh Filters

Scrubbing systems involving a wetted-knit mesh filter have recently been
developed and promoted by Heat-Systems—Ultrasonics and Dupont.1^!15'
In both systems the gas stream is first sprayed with water and then the
gas-liquid mixture is passed through a fiber bed structure having a pore
size sufficiently fine and a path length sufficiently long to provide
intimate contact of the gas stream containing the particulate solids and
mist with the liquid.
                                    42

-------
Figure 3 and Table 5 present some performance data reported by Dupont
for a laboratory-scale system.ll/  The characteristics of the bed used
to obtain the data were:

1.  Bed diameter--10.26 cm (4 in.)

2.  Fiber material--304 stainless steel in knitted form

3.  Packing density—128 kg/m3 (8 lb/ft3)

4.  Bed void fraction--0.983

5.  Specific surface area 459 n^/m3 (140 ft2/ft3) for 6-mil fiber diam-
eter and 754 n^/m3 (230 ft2/ft3)

Talc dust with the following size distribution was used in the tests:
50%  < 2.4 microns, 287. < 1 micron, and 14% < 0.5 micron.  The test re-
sults in Table 5 show quite high collection efficiencies with relatively
low pressure drop.  Sampling procedures used to obtain the data in Table
5 were not reported, but assuming that valid procedures were utilized,
quite high collection efficiencies appear possible with relatively low
pressure drop.  If the efficiencies reported by Dupont for their laboratory-
scale system can be substantiated by testing of pilot-scale or demonstra-
tion-scale systems, wetted-knit mesh filters will be an addition to tech-
nology for the control of fine particulates.

PROPOSED CONTROL SYSTEMS

Numerous concepts have been advanced for the control of particulates over
the years.  Investigation of the proposed systems has been random in
nature and the potential of the proposed systems is generally an unknown.
Several broad categories of proposed systems (see Table 4) were analyzed
in this program in order to define their potential for use in controlling
fine particulates.  Appendices C-I present detailed discussions of the
systems evaluated.  The principal observations for each category are
presented next.

Condensation Scrubbers

Condensation scrubbing involves particle removal from the gas stream via
the use of a temperature gradient, a vapor concentration gradient, vapor
condensation, or a combination of the three.  It is not necessary that
all forces act simultaneously, but in most actual cases they do.
                                   43

-------
o
I/I
  in
  o
 o
 1
      10,000
1,000
         100
          10
         1.0
                    3.5 Mil Fiber
                    Diameter
                                             6 Mil Fiber
                                             Diameter
  99.99
  99.9
                                                     99.0
                                                      90.0
                                                      0%
                                            8
                         THICKNESS OF FIBROUS
                         BED STRUCTURE - INCHES
                                            10
12
 Figure 3.  Measured efficiency of Dupont wetted-knit mesh separator
                        collecting 2 micron particles!!/
                                                            U

                                                            LLJ
                                                            U
                                                                   O
                                                            d
                                                            U
                                   44

-------
                                                                   Table 5.  TEST DATA ON DUPONT SCRUBBER
in

Run
No.
1
2
3
4
5
6
7
a
9
10
11
12
13
14
15
16

Bed
(ca)
10.2
10.2
10.2
10.2
10.2
10.2
10.2
10.2
20.3
20.3
10.2
10.2
10.2
10.2
10.2
10.2
Liquid Spray Rate
(liters/
min-m2 (gpm/fc-
Deoth Bed Density of of
(In.) (kg/ml) (lb/ft3) Bed Material chamber) chamber)
4 64.1 6 ( Bed 896 22
4 composed
4 of
4 16 nil
4 diameter
4 4 knitted 977 24
4 polyethylene
4
8
8
4 320.4 20 304 atalnleia ateel 407 10
4 16-mll diameter
4 640.7 40 304 stainless 407 10
4 ' 11-mll
4 192.2 12 diameter 407 10
4


Gas ThrouEhout
m ,'mln)
5.04
5.89
1.30
1.84
2.27
1.36
2.66
5.04
1.36
2.61
3.71
5.10
4.11
5.01
3.88
4.96
(ftj/mln)
178
208
46
65
80
48
94
178
48
92
131
180
145
177
137
175


Superficial
Gas Velocity
(m/sec)
10.4
12.2
2.7
3.7
I..-J
2.7
5.5
10.1
2.7
5.5
2.7
3.7
3.0
3.7
2.7
3.7
(ft/secl
34
40
4
12
16
9
18
33
9
18
9
12
10
12
9
12


Inlet Dust
Concentration
(grains/
30
32
11
J5
37
64
43
32
149
110
85
.,
b4
82
50
62
13
14
5
11
16
28
19
14
65
48
37
37
28
36
22
27


Pressure Drop
(cm IliO)
28.7
34.8
7.4
8.6
11.4
10.9
20.6
33.5
22.9
48.3
9.4
12.2
24.4
25.1
27.2
26.2
(In. H20)
11.3
13.7
2.9
3.4
4.5
4.3
8.1
13.2
9.0
19.0
3.7
4.8
9.6
9.9
10.7
10.3

Collection
Efficiency
98.9
98.7
97.7
98.4
99.2
99.8
98.8
98.9
99.4
99.6
98.8
98.5
98.6
98.9
99.4
98.7

-------
The work on condensation scrubbing conducted to date indicates that the
collection of submicron particulates from various gas streams can be en-
hanced by the use of condensation effects.  The collection efficiency is
a function of;

1.  Amount of water vapor condensed per unit mass of dry gas;

2.  Particle concentration with a trend of increasing efficiency with
decreasing particle number concentration; and

3.  Type of particle being scrubbed (efficiency varies in the order hy-
groscopic dust > hydrophillic dust > hydrophobic dust).

Our analysis (see Appendix C) of the potential of condensation scrubbing
for particulate control indicates that it can be an important avenue to
exploit in order to collect fine particulates.  Optimization of condensa-
tion effects in venturi and orifice type scrubbers could result in de-
creased power consumption for those classes of conventional wet scrubbers.*
Condensation phenomena can also provide the basis for new families of con-
trol systems for fine particles.

Two avenues that could be exploited to develop new control systems are:
(a) injection of steam into a cold dust-laden gas, and (b) sudden cooling
of a hot and humid gas stream.  Under proper thermodynamic conditions,
injection of steam into a cold dust-laden gas stream can result in the
conversion of a portion of the steam into condensed droplets.  Water
vapor would preferentially condense on the dust and, thereby, increase
the effective particle size of particulates} making collection of the par-
ticulates easier.  Treatment of gas streams by steam injection would not
be very effective for gas temperatures above 100-120°F because condensa-
tion would not be significant.  Sudden cooling is an inefficient method
for producing condensation since most of the water will condense on the
cooling surface.  However, enough condensation will occur on the particles
to cause some particle growth.

Multiple stage or continuous-type contact equipment is superior to single-
stage equipment for both injection and cooling.  Figures C-2 and C-3,
Appendix C, present possible equipment configurations.  Distribution of
condensation over several stages is preferable because of the enhanced
growth which can occur after the particle concentration has been reduced.
   Standord Research Institute is currently conducting research in this
     area under funding from the Control Systems Laboratory.—'
                                   46

-------
 The  economics  of  condensation will be dependent upon  the source and  the
 characteristics of  the emitted particulates.  Costs for condensation
 scrubbing will be highly dependent on the amount of steam condensed,
 especially if steam or fuel to evaporate water must be purchased.  Cooling
 water may also be required to condense the steam and  the amount of cooling
 water needed will depend on the temperature rise of the water in the
-scrubbing system.

 Currently, limited  information is available on the details of the phenomena
 occurring when steam, water droplets, and particulates are mixed--espe-
 cially in equipment configurations that are likely to be useful in actual
 industrial gas cleaning applications.  In order to resolve the areas of
 uncertainty and to  develop reliable data for engineering design, additional
 development effort  is necessary.  Effort should be focused on pilot-scale
 equipment in order  to develop data under realistic heat and mass transfer
 conditions.  The  Control Systems Laboratory is currently funding work
 in this  area at A.P.T., Inc., Riverside, Calif ornia.lili/

 Charged  Droplet Scrubbers

 Control  devices using charged drops as the collecting bodies are currently
 under  study by several investigators JJil!2/  Currently available data on
 charged  droplet scrubbers and related devices indicate some potential
 for the  collection  of fine particles.  However, adequate data are not
 now available  for a comprehensive analysis of any of  the devices.  While
 the ultimate efficiency and usefulness of these devices can not be deter-
 mined  accurately  at this time, the major workers in this field estimate
 from their experience and analysis that efficiencies  comparable to,  but
 not much better than, conventional electrostatic precipitators can be
 obtained for submicron particles.18!20"21/

 Potential advantages of charged droplet systems in comparison to con-
 ventional electrostatic precipitators include the ability to clean gases
 at low voltage levels and elimination of problems of  dust reentrainment
 and of dust resistivity frequently encountered in conventional electro-
 static precipitators.  Advantages with regard to wet  scrubbers include
 lower  water usage and lower pressure drop at the same collection effi-
 ciency.  Disadvantages in comparison with conventional electrostatic
 precipitators  include:
                                    47

-------
 1.  Need  to produce and charge drops;

 2.  Premature space charge precipitation of drops on walls of device;

 3.  Neutralization of charge of drop upon collection of oppositely
 charged particulates; and

 4.  Need  to handle wastewater stream.

 The main  application areas for charged droplet scrubbers are likely to be
 on sources where conventional electrostatic precipitators or wet scrubbers
 do not perform satisfactorily.  Total energy requirements for charged
 droplet systems are relatively uncertain, but about 0.07 hp/m^/min
 (2 hp/1,000 cfm) appears to be a reasonable estimate.

 It is not clear whether the performance of charged droplet systems will
 be sufficiently greater than alternative control systems to justify their
 development and subsequent use on industrial gas cleaning applications.
 Additional research and development activity on charged droplet systems
will be required to determine their real merit.  Emphasis on future R&D
 should be placed on experimental work on bench-scale or pilot-scale model
 systems so that meaningful engineering design data are obtained.  Work
 is currently under way at TRW, Inc., and the University of Washington
 on charged droplet systems.  The former work is being supported by
 EPA/CSL, while the latter is supported by industry.

 Foam Scrubbers

The addition of surface-active agents to water produces a foam and the
 effectiveness of the collection of particulates by bubbles in wet scrub-
bers may be increased by the use of foam.  At present, the mechanism of
 foam scrubbing is not well-defined and the utility of foam scrubbing is
difficult to determine.

The limited experimental data on foam scrubbers reported in the technical
 literature indicates good results for aerosol collection by foam.23"2^/
Reference 25 describes a foam bed scrubber which is a product of a Swedish
 firm.  Reference 25 reports that the foam scrubber has been tested on a
variety of dusts and that collection efficiencies in the 90-977. range are
obtained.  Information on fractional efficiency of the foam scrubber was
not given in Ref. 25.

Based on the limited available information,  the potential of foam scrub-
bers to control fine particles would seem to be good if the residence
                                   48

-------
time is sufficient to collect fine particles.  Potential problems in foam
scrubbers are formation of fine bubbles for good collection, foam destruc-
tion and secondary foaming.  More detailed information is required re-
garding the mechanism of bubble formation, splitting, and destruction
before the technical and economic feasibility can be determined.

Electrified Filters

One potential technique for obtaining improved particulate control de-
vices is to use electric fields in conjunction with various filter sys-
tems.  Fabric filters, wire-mesh filters, granular beds, or fluid beds
can be used as the underlying filter elements.  The magnitude of elec-
trical forces compared to other forces that can be applied to particulates
suggests that improvement in aerosol filtration should occur when electric
fields are present in conjunction with filter systems.

Limited theoretical and experimental studies on idealized systems indi-
cate that electric fields do improve the collection efficiency of filter
systems.  However, it is not clear whether the improved performance of
these filters is sufficient to justify their use or whether electrical
filters will be adaptable to the control of emissions from industrial
processes.  Research to explore the feasibility of this class of devices
in more detail is warranted.

Granular Bed Filters

The ability of porous beds of granular materials such as sand, coke,
sintered materials, Raschig rings, or similar materials to remove par-
ticles from gas streams has been known for many years.  Although granular
beds have found practical application in atomic energy facilities and the
filtering of small volume gas streams, application to industrial sources
of particulates has been limited.

At present, there are no useful models for aerosol filtration in granular
beds.  Knowledge of the performance of granular beds has been obtained
essentially from experimental studies at both the laboratory and pilot-
scale level.  Experimental investigations of granular beds have been per-
formed on fixed and moving beds and beds augmented by electrostatic and
sonic forces.  Meager results on pilot-scale granular bed filters operating
on industrial sources of particulate pollutants indicate collection effi-
ciencies in the range 85-987..—'  Reference 27 reports that a 5.1 cm
(2 in.) thick electrified granular bed collected 0.6-0.7 micron methylene
blue aerosol with an efficiency of 99.5% at a pressure drop of 0.51 cm
wg.
                                    49

-------
Fluidized Beds

Fluidized beds have also been proposed as particulate filtration media.
Experimental studies on fluid beds have shown that collection efficiencies
of 90-937. can be obtained from single pass fluid beds.  Attainment of high
efficiencies by staging may be possible even though single stage efficiencies
are not high.

Electrostatically augmented fluidized beds have been studied at Harvard
University?!/ and MIT.15/  Experimental studies at Harvard University
showed collection efficiencies in excess of 90%.

With the possible exception of electrostatically augmented systems, fluid
beds do not offer sufficient potential for fine particle control to war-
rant additional research.  In view of other more promising and simpler
approaches, the electrostatically augmented system is also of questionable
utility.

Gamma-Ray Precipitator

The idea of utilizing ionizing radiation to charge particulate matter so
that it can be collected by electric fields in an electrostatic precipi-
tator has been proposed in the patent literature.29^32/  Experimental
studies of the gamma-ray precipitator concept have recently been con-
ducted at Penn State University (see Appendix I).  The results of limited
tests on the emission stream from a small stoker-fed home-type, furnace
indicated collection efficiencies comparable to those of conventional
electrostatic precipitators.

Potential advantages for the gamma-ray precipitator in comparison to con-
ventional electrostatic precipitators are:  (1) possible higher ion
densities; (2) lower operating costs; and (3) simplified cleaning sys-
tems.

The major disadvantages are the radioactive material required and the
shielding structures.  Both of these disadvantages will greatly limit
the use of such a device even if it is proven highly efficient.  Utility
companies with experience in handling nuclear materials would appear to
be the only likely applications for such a device.
*  The work at MIT has been supported by the Control Systems Laboratory.
                                   50

-------
Sonic Agglomerators

Sonic agglomerators can be an effective way of agglomerating fine par-
ticles  to  a  larger size which could be collected by conventional control
equipment.

Sonic coagulation of particulates has as its principal advantage its
applicability  to any aerosol, including those comprised of submicron
particles.   The principal disadvantage of sonic coagulation is the rela-
tively  high  energy requirements, which have been reported to range from
0.05-0.7 hp/m3/min (1.5-19 hp/1,000 cfm).  Mednikovl/ suggested that
minimum total  energy requirements of 0.16-0.32 hp/m3/min (4.5-9 hp/1,000
cfm) should  be achievable with optimum design.  Presumably this energy
requirement  is based on agglomeration to a sufficient size to be removed
by  inertial  separators.  A second major disadvantage is the low effi-
ciency  of  acoustic coagulators and their inability to handle highly dis-
persed  suspensions.

The energy requirements of a sonic precipitator with an inertial separator
or  other simple filtering system are relatively high.  Water augmenta-
tion appears necessary to obtain the removal of a large fraction of the
particulates.  If 0.02 kg of water per cubic meters per minute is used,
one would expect an increase in the power requirements of up to 0.04 hp/
nr/min  depending on the device used and the mean droplet size.  The
minimum energy requirements would then be 0.2-0.35 hp/m3/min (5.5-10 hp/
1,000 cfm)

Alternatively, one should consider hybrid units.  For example, sonic
agglomeration  could be used upstream from a conventional electrostatic
precipitator.  Since the particulates now need to be agglomerated to only
1-2 microns instead of 10-15 microns with an inertial separator, a re-
duction in acoustic energy requirements by at least a factor of 5 would be
expected.  Other hybrid units should also be possible.

SUMMARY

In the category of emerging and proposed control systems for fine par-
ticulates, there are several systems that hold promise for transition to
commercially viable systems.  Foremost in the category of emerging sys-
tems are the Steam-Hydro Air Cleaning and the ADTEC wet scrubber systems.
Both of these systems look especially attractive for use on sources where
sufficient waste heat is available to provide energy requirements.  Fur-
naces used to process various metallurgical materials would appear to be
prime sources for application.
                                    51

-------
Only limited data are available to judge the potential of most of the
proposed systems.  Assessments must be based predominately on the ability
of the underlying collection forces or mechanisms to collect fine par-
ticulates.  Condensation scrubbers, charged droplet scrubbers, elec-
trified filters of various types, and foam scrubbers have sufficient
potential to merit additional research and development.  The Control
Systems Laboratory is funding research programs in some of these areas
at the current time and a more extensive program appears warranted.
Since in general only data on laboratory-scale equipment are available
for proposed systems, future studies should be conducted in equipment
configurations and sizes that will produce reliable engineering design
data.
                                   52

-------
                             REFERENCES

 1.  Goldsmith, P., H. J. Delafield,  and L.  C.  Cox,  "The Role of Diffu-
       siophoresis in the Scavenging  of Radioactive  Particles from the
       Atmosphere," Q. Jl R. Met.  Soc.. 89,  43  (1963).

 2.  Goldsmith, P., and F. G. May, "Diffusiophoresis and Thermophoresis
       in Water Vapor Systems," Chapter 7 of Aerosol Science, pp.  163-
       194, C. N. Davies, Ed., Academic Press,New York  (1966).

 3.  Sparks, L. E., and M. J. Pilat,  "Effect of Diffusiophoresis on
       Particle Collection by Wet  Scrubbers," Atmospheric Environment.
       4, 651-660 (1970).

 4.  Calvert, S., et al., "Feasibility of Flux  Force/Condensation Scrub-
       bing for Fine Particulate Collection," EPA Report EPA-650/2-73-
       036, October 1973.

 5.  Mednikov, E. P., Acoustic Coagulation and  Precipitation  of  Aerosols.
       Pergamon Press (1964).

 6.  Fuchs, N. A., The Mechanics of Aerosols. Pergamon  Press  (1964).

 7.  Zebel, G., "Coagulation of Aerosols," in Aerosol Science. C.  N.
       Davies, Ed., Academic Press, New York (1966).

 8.  Oglesby, S., Jr., and G. B. Nichols,  "A Manual  of  Electrostatic
       Precipitator Technology,  Parts  I and  II,"  Southern Research
       Institute, Birmingham, Alabama  (1971).

 9.  Private communication,  Mr.  J. A.  Abbott, Environmental Protection
       Agency, Control Systems Laboratory, September  1973.

10.  Darby, K., and K. R. Parker,  "The Electrostatic  Precipitator  and
       Its Range of Application,"  Filtration and  Separation. May-June
       1972.
                                  53

-------
11.  Calvert, S., et al., Scrubber Handbook (EPA Contract No. CPA-70-95),
       APT, Inc., Riverside, California, August 1972.

12.  Private communication, Mr. T. W. Evans,  Lone Star Steel, December
       1973.

13.  "A New Process for Cleaning and Pumping  Industrial Gases -  The
       Aronetics  System," U.S. Patent No.  3,613,333.

14.  Sales Brochure, "Heat-Systems—Ultrasonics," Plainview,  New York.

15.  Lucas, R. L., and H. F. Porter, "Process and Apparatus for  Wet
       Scrub Removal of Dust and Mist from  Gases," U.S. Patent No.
       3,370,401.

16.  Kostow, L. P., "Design^ and Testing of  Space-Charge Precipitators,"
       M.S. Thesis, University of California, Berkeley, California,
       March 1972.

17.  Melcher, J. R., and K. S. Sachar, "Charged Droplet Technology for
       Removal of Particulates from Industrial Gases," EPA Contract No.
       68-002-0018, MIT, Cambridge, Massashusetts (1971).

18.  Private communication, Dr. Michael Pilat, University of  Washington,
       September 1973.

19.  Private communication, Mr. B. Koppang, TRW, Inc., September 1973.

20.  Melcher, J. R., "Progress Report - Research on Systems of Charged
       Droplets and Electric Fields for the Removal of Submicron Par-
       ticulates from Industrial Gases," MIT  Report, EPA Contract No.
       68-002-0018, April 1973.

21.  Private communication, Dr. J. R. Melcher, MIT, September 1973.

22.  Luchnskii, G., Zhurnal Fig Khim. 13, 302 (1939).

23.  Alekseeva, B., and M. Andronov, Lab Prakt USSR. ^6(1), .18 (1941).

24.  Bronsiky, D., and F. Divoky, Chem. Met Engr.. 47, 541 (1940).

25.  Javorsky, B. S., "Gas Cleaning With the  Foam Scrubber,"  Filtration
       and Separation, March-April 1972.
                                   54

-------
26.  Kalen, B., and F. A.  Zeng,  "Filtering Effluent  from a  Cat  Cracker,"
       Chemical Engineering Progress.  6£(6),  67  (1973).

27.  Cole, W. H., "Electrical Precipitation and  Charged  Particle  Col-
       lecting Structure Therefore," U.S.  Patent No.  2,990,912, 4 July
       1961.

28.  Juvinall, R. A., et al., "Sand-Bed Filtration of Aerosols:   A Review
       of Published Information on Their Use  in  Industrial  and  Atomic
       Energy Facilities," Argonne National Laboratory Report ANL-7683,
       June 1970.

29.  McGray, H. F., U.S. Patent No. 1,991,934.

30.  Jacob, C. W., U.S. Patent No. 2,381,455.

31.  Leupi, R., et al., U.S. Patent No. 2,934,648.

32.  Schmidt, C. M., U.S.  Patent Application  No. S.N. 83,874, October
       1970.

33.  McCain, J., "Performance of Electrostatic Precipitators,"  presented
       at EPA Seminar Electrostatics  and Fine Particles. Research Tri-
       angle Park, North Carolina, September  1973.

34.  Billings, C. E., and  J. Welder,  Handbook of Fabric  Filter  Technology.
       Volume I. GCA Corporation Report GCA-TR-70-17-G,  EPA Contract No.
       CPA-22-69-38, December 1970.

35.  "Wet Electrostatic Precipitator Tames Oils, Gases,  and Fines,"
       Chemical Engineering, pp. 74-76, July  1973.

36.  Calvert, S., et al.,  "Feasibility of  Flux Force/Condensation
       Scrubbing for Fine  Particulate  Collection," EPA Report EPA-650/
       2-73-036  (EPA Technology Series).
                                   55

-------
            GLOSSARY OF TERMS,  ABBREVIATIONS,  AND SYMBOLS




 ai = thermal reflection coefficient




  A = area of collecting surface




  B = magnetic field strength




 Cp = specific heat of gas




  C = Cunningham correction factor




  D - diffusion coefficient




 Dc = diameter of collecting body




DBM ~ particle diffusion coefficient




 Dp = particle diameter




  d = distance between particles




  E = electric field strength




  f = frequency of vibration




  F = force




  g = gravitational constant




  h = bed height/bed diameter




 kg = thermal conductivity of  the gas




 kp = thermal conductivity of  the particle




  k = Boltzmann constant






                                    56

-------
   L = length


  m^ = mass of molecules of component  i



  nip = mass of particle



  n^ = number of molecules of component  i  per unit volume



  Nj = inertial impaction parameter



  NQ = ion density



  NR = interception parameter




 %e = Iteyrtold'8 number


 Npe = Peclet number



  P! = partial pressure of component  i



   p = pressure



   F = penetration



  q  = electrostatic charge on particle
   P


  qf = electrostatic charge on collector



  r  = particle radius



   r = radius of rotation



  Rb = radius of bubble



   T = temperature



   t = time



v    = particle velocity relative to gas
 r o


v    = particle velocity relative to collector



  VT = thermophoretic velocity (particle)



  v_ = diffusiophoretic velocity (particle)



  v  - particle terminal drift velocity in magnetic field



   v = particle velocity

-------
vtp »  tangential velocity of particle




vtg = tangential velocity of gas




 Vjj = bubble velocity




 v  = linear gas velocity
  5



  V = volume flow rate of gas




  W = width of chamber




  # = coefficient of absorption




  g = volume shape factor




  e = rate of dissipation of turbulent energy




 e  = dielectric constant of particle




 6£ = dielectric constant of gas




 CQ = permittivity of free space




  \ = mean free path of gas




 \s = wavelength of sonic vibration




 Xo = internal scale of turbulence




  $ = particle mobility




  71 = collection efficiency




  6 = kinematic viscosity




 Pc = coagulation constant




 p  = particle density




 p  = density of gas




  u) = migration velocity in electric field




  u = gas  viscosity                  _Q
                                     JO

-------
                             APPENDIX A

               COLLECTION AND AGGLOMERATION MECHANISMS
                        FOR FINE PARTICULARS
INTRODUCTION

Farticulate collection is effected by passing a gas stream through a
system where the particles are acted on by forces which remove them from
the gas stream.  The forces acting give the particles a component of
velocity in a direction other than that of the gas stream.  To be effec-
tive, these forces must be sufficiently large to take the particle out
of the gas stream during its residence time in the system.  If the par-
ticulates in the gas stream are submicron in size, their removal may be
facilitated by making agglomerates of the very small particulates and
then collecting the agglomerates.

The basic mechanisms or forces that can be used to collect or agglomerate
particulates are shown in Table A-l.  The effectiveness of the individual
mechanisms is strongly dependent upon factors such as particle size, flow
velocity, particle density, particle number density, temperature gradients,
concentration gradients, and heat and mass transfer coefficients.  As a
result, many of the forces or mechanisms in Table A-l are not effective
for the collection of fine particles.  The important parametric dependence
of each force or collection mechanism is discussed in the following sub-
sections.

PARTICULATE COLLECTION

Gravity and Momentum Forces

In gravity separators, the gas stream is slowed down sufficiently to
allow particles to settle out.  The diameter of the smallest size particle
that is completely settled out is given by Eq. (A-l)
                                    59

-------
          Table A-l.  PARTICLE COLLECTION OR AGGLOMERATION
                         FORCES (MECHANISMS)
I.  Particle Collection

     A.  Gravity and momentum forces
     B.  Centrifugal forces
     C.  Aerodynamic capture

          1.  Inertial impaction
          2.  Interception
          3.  Diffusion
          4.  Electrostatic attraction
          5.  Gravitational settling

     D.  Flux forces

          1.  Electrostatic forces
          2.  Thermal forces
          3.  Diffusion forces
          4.  Magnetic forces

II.  Agglomeration and/or Particle Growth

     A.  Thermal or Brovnian agglomeration
     B.  Turbulent agglomeration
     C.  Electrostatic agglomeration
     D.  Sonic agglomeration
     E.  Condensation
                                     60

-------
                                            !/2                /* -.N
                          D   -i 18uV                          (A-l)
                           P"   |LW(pp-pg)g
                                ^          •

where    Dpm = diameter of the smallest size particle that is
                 completely settled out,

          p  = particle density,

          p  = fluid density,

           g = local gravitational constant,

           u = absolute viscosity,

           L = length of chamber,

           V = volumetric flow rate of the gas at the actual
                 temperature and pressure, and

           W = width of chamber.

Examination of Eq. (A-l) shows  that in theory a very large settling
chamber would give sufficient time for even small particles to settle
out.  In practice, space limitations restrict the use of settling chambers
to large particles.

The efficiency of a simple settling chamber can be improved by giving
particles a downward momentum in addition to the gravity settling ef-
fect.  The number of possible designs incorporating this principle is
very great, varying from a simple baffle in the chamber to specially de-
signed jets which give accelerated settling.  While momentum collectors
do improve overall collection efficiency, they do not significantly en-
hance the collection of fine particles.
                                    61

-------
Centrifugal Force

Particulate matter is separated from gas in a cyclone by centrifugal
force, or radial force, tending to drive the particles (against the
resistance of motion by the gas) to the cyclone wall.  The radial force
imparted to the particle is
                                                               (A-2)
where   Fs = radial separating force,

        mp = particle mass,

       vtp = tangential velocity of particle,

         g = gravitational constant, and

         r = radius of rotation.

The tangential velocity of the particle is theoretically lower than that
of the gas in which it is suspended; but, in considering fine particles,
the two velocities are equal for all practical purposes and  vtp = vtg  .

Equation A-2 can also be written in the form



                                             •                 (A'3>
where   0 = volume shape factor, dimensionless,

       pp = particle density,

       Dp = particle diameter,

       rj = — (ratio of radius of rotation to the radius of confining
             s   structure), and

        n = exponent, dimensionless

The force resisting particle motion is given by Stokes1 law
                                  62

-------
                          Fr = CnDpvp_g   ,                     (A-4)

where  Fr = frictional resistance to flow,

        C = Cunningham factor,

     vp-g = particle velocity with respect to gas, and

        ji = gas viscosity.

The ratio of the separating force to the  resisting force
                           Is =     pp                       (A.6)
                           Fr   Cguvp_grr£2n

provides an indication of  the relative ability to remove various size par-
ticles in a cyclone.  Since  the force ratio varies with  Dp2  ,  large
diameter particles are more  readily separated in a cyclone.   Commercial
cyclones are effective in  collecting particles down to about  10 microns,
and carefully designed units can collect particles down to about 2-5
microns .

Aerodynamic Capture

Aerodynamic capture of particles involves  the collection of particles by
collecting bodies  (e.g., fibers, packing,  droplets, etc.).  To  utilize
aerodynamic capture,  the gas stream is brought near the collecting bodies
and then a number  of  short-range mechanisms accomplish the actual  collec-
tion.  The most effective  mechanisms are:   inertial impaction,  intercep-
tion, diffusion, and  electrostatic  attraction.  The relative  importance
of each mechanism  varies with  the size and velocity of the particles and
the collecting body, itself .

Inertial Impaction

The effectiveness  of  inertial  impaction  is a function of  the  inertial
impaction parameter which  arises out of  the force balance equations of
fluid resistance opposing  the  motion of  the particle. The inertial im-
paction parameter  is  defined as

                                                                (A.6)
                                      Dc
                                     63

-------
where  C  is Cunningham correction factor for Stokes1 law, v  is particle
velocity downstream, pp  is particle density and  p  is fluid viscosity,
Dp = particle diameter, and  Dc = diameter of collecting body.

Ranz, Wong and Johns ton&i' and other investigators have shown that the col-
lection efficiency of inertial impaction is a function of the impact ion
parameter and that the impaction parameter (Nj) must exceed 4 for collec-
tion efficiencies to exceed 90%.  For a specific particulate system,
Cpp/18u will be a constant and to achieve a high collection efficiency of
fine particles, only the particle velocity and the diameter of the collecting
body can be varied to increase collection efficiency.  Assuming unit density
particles and typical gas viscosities, the necessary ratio of v/Dc  re-
quired to result in a value of  Nj  exceeding 4 is shown in Figure A-l as
a function of particle size.  The cgs unit system was used in deriving the
data shown in Figure A-l.  Figure A-2 shows the particle velocity required
for high collection as a function of the diameter of the collecting body
and the diameter of the particle to be collected.  It is apparent from
Figure A-2 that in order to collect submicron particles efficiently by
inertial impaction, it is necessary to use high velocities and small col-
lecting bodies.

Interception

Whenever the streamline along which the particle approaches a collecting
body passes within a distance of one-half the particle diameter from the
body, interception of the particle by the collecting body will occur.
This mechanism would never occur alone except as a limiting' case for
particles of low density.  However, it should be taken into account as a
boundary condition to be met along with other aerodynamic capture mechanisms.

Using Langmuir's viscous flow equation, Ranz!' obtained the interception
collection efficiency of a cylindrical target
                  E%  (2+NR)~|
!  (1+NR) In (1+%)   (1+%)   I
                       !  (1+NR) In  (1+%)                        (A-7)
where   A =	 , and
            2  (2.002-In NRe)
            Dc
                                   64

-------
                                         ASSUMPTIONS:
/>p=1.0g/cc

 i = 1.8x 10~4
                                                      po'se
                                          v = cm/sec

                                            =cm
                    PARTICLE DIAMETER - MICRONS

Figure A-l.   Ratio of particle velocity to collector body diameter
                 necessary to achieve 907. target efficiency
                           for inertial impaction
                                   65

-------
     107
 u
 
t
 u
u
9
$
a
     105
     104
     103
         1
                  100
                                                  Dp =0.1 Micron
                                                 Dp =0.5 Micron
                                                  Dp = 1 Micron
                                     ASSUMPTIONS:


                                     />p = l.Og/cc


                                     /x = 1.8 x 10"4 poise
200
300
400
500
              DIAMETER OF COLLECTING BODY (Dc)- MICRONS

     Figure A-2.   Particle velocities  required to achieve  90% target

                          efficiency  for inertial impaction
                                    66

-------
This equation should be used instead of  the more simplified equations  re-
ported  in some references because it allows for changing flow patterns
with different stream velocities.  Friedlanderl/ found  that a log- log  plot

of - ,  as calculated from this equation, against  NR  was a straight line
which could be represented by
                             = 1.25 NR'                        (A-8)

in the range of 0.05  < NR < 1 .

Figure A- 3 presents the interception parameter as a function of  N^e  for
target efficiencies of 0.5 and 0.9.  Inspection of Figure A- 3 shows that
for interception to be a significant mechanism for particle collection,
the diameter of the collecting body must be of the same order of size as
the particles to be collected.  In addition, the flow velocity must be
low.  These restrictions imply that interception will not play a signifi-
cant role in fine particle collection in most practical situations unless
multiple targets are provided.

Brownian Diffusion

For very small particles, Brownian movement will be superimposed upon the
flow motion of a particle.  This relatively slow diffusional velocity may
be sufficient to cause the particle to come into contact with the col-
lector if the particle passes close enough to the collecting body and re-
mains there for a long enough time.

Researchers working in this area have developed expressions for the target
efficiencies due to Brownian diffusion which indicate that the target ef-
ficiency is an inverse function of the Peclet number.  The Peclet number
is defined by Eq. (A- 9)
                             NPe = --   ,                     (A-9)
                                   DBM

where  DBM = particle diffusion coefficient.

Since the diffusion coefficient is an inverse function of the particle
size, the Peclet number can also be written as
                           NPe%DcDP   •                      (A-10)

                                   67

-------
       3.0
       2.0
*l
  II
  a
 Z
 u
a
       1.0
               0.1
                                    0.5
                                       N
                                        Re
1.0
     Figure A-3.  Interception parameter as a function of Reynolds

                      Number for selected target efficiencies
                                      68

-------
Figure A-4 presents graphical solutions to the equations for  target  ef-
ficiency developed by Ranz,— ' Friedlanderr' and Langmiur.— '   Inspection
of Figure A-4 indicates that the collection efficiency attributable  to
this mechanism is not significant if the value of  Npe  is much above 10,
In most situations, particle collection by Brownian diffusion will be of
importance vonly for particles less than 0.5 micron in diameter.

Electrostatic Attraction

When an aerosol particle, or a stationary object in a flow stream, is
electrically charged, or when both the particle and the object carry
electric charges, the trajectories of the aerosol particle past the  ob-
ject are affected usually with an increase in the number of particles
colliding with the object.

Ranz and Wong—' and Kraemer and Johns tone-' defined the dimensionless
force ratios given by Eqs. (A- 11) and (A- 12) to characterize  the forces
between an aerosol particle in the absence of a field across  a filter.


                                 "*                         (A-ii)
                                3nue0DpVp_c
                                                                (A-12)
where   ep  and  ef  are dielectric constants of  the particle and  the  gas
and  q   and  qf  are electrostatic charges for  the particle and  the  col-
lector,,  eQ  is permittivity of free space and   u  is viscosity of  the  air.
Equation (A-ll) describes  the interaction of a charged particle and col-
lector, and Eq. (A-12) describes  the interaction between a charged  col-
lector  and a dielectric particle  on which the collector induces a charge.
Parameters  Kg  and  Kj  may be considered ratios of  the electric  force
at the surface of  the collector to  the fluid resistance caused by  a  rela-
tive particle velocity of vp_c with respect to  the collector.  It  is
noted that when  q  and  qf  are of  the same sign, Kg  is positive  and
collection efficiency decreases.  Target efficiency by these mechanisms
is negligible when the corresponding  parameter  is much less than 10"2
and is of the order of unity when the parameter is of the order  of unity.
                                    69

-------
                                          B jFriedlander
      1(T3
         10'
10
         Npe- DIFFUSION PARAMETER - PECLET NUMBER
Figure A-4.   Target efficiency by diffusion on an isolated cylinder
                                70

-------
Kraemer and Johnstone^' suggest that target efficiencies can be expressed
for induced electrostatic attraction by

                                        1.4
                                                               (A-13)


and for charged particle and collector electrostatic attraction by


                              T] « - 4 KE                       (A-14)

based upon experimental data for a spherical collector.

COMPARISON OF AERODYNAMIC CAPTURE FORCES

Although the forces of the various aerodynamic capture mechanisms are
additive, the resulting individual efficiencies are not directly additive.
Each mechanism will contribute to a total efficiency, however, and no com-
bination of favorable mechanisms will cause an efficiency lower than that
expected for any one of the favorable mechanisms.  In addition, the effi-
ciency of collection for any given mechanism will be proportional to the
size of  the particle and the size of the collector, and to  the particle-
collector differential velocity in approximately the same manner that the
parameter characterizing that mechanism is proportional to  Dp  , Dc  , and
Vp-c  • by definition.
The approximate target collection efficiencies for spheres  calculated by
the foregoing equations are shown in Figure A-5.  Inspection of Figure
A-5 shows that single  target efficiencies for electrical forces are  gen-
erally larger than  those for impaction,  interception and diffusion  at
low velocities.

FLUX FORCES

Particles can be collected by forces which result from electrical, tem-
perature, and concentration gradients,  from the flux of matter or energy,
and from a magnetic field.  This group  of forces, often termed flux  forces,
are attractive for  the collection of fine particles, because the magnitude
of the flux forces  does not approach zero as the size of the particles  to
be collected approaches  the submicron range.  Individual flux forces are
discussed  in the  following sections.
                                    71

-------
         ioor
          10
     I
     £
     LU
        10
          -1
        10
        10'
r2
                            K
 10"6   ID'5   10'4   10-3   ID'2   10'1     1      10

         Collection Parameter,Nj, NR,  Npe, K], KE
                                                                 100
Figure A-5.  Approximate solutions of equations for target efficiencies
                                    72

-------
Electrical Forces (Electrophoresis)

If charged particles are subjected to an unidirectional electric field,
they move towards the electrodes and are deposited.  The motion or
migration of the particles in the field is termed electrophoresis.  In
the absence of turbulence or other aerodynamic effects, the migration
velocity  <*>  resulting from the electrostatic force can be obtained from
Stokes1 law and is given by
                                u> = -SE—                       (A-15)
                                    6TTrpii

where  q  is the charge on  the particle, E  is  the strength of  the  elec-
tric field, r_  is the particle radius and  u   is the viscosity of  the
gas.  This expression neglects second order electrostatic  effects such
as  the polarizability of  the particle, and assumes that  a  spherical par-
ticle is moving in laminar  flow (NRe < 1).  Both of  these  assumptions
are normally adequate.!/  It is clear from Eq.  (A-15) that for  a given
particle size the only practical variables affecting the migration  velocity
are the electric field strength and the  charge.

To  secure rapid and  efficient deposition of particles, forced charging  of
particles may be adopted  and this forms  the basis of electrostatic  pre-
cipitators.  The electric field strength is determined by  the configuration
of  the specific control device.  The charge on a particle  is a  very strong
function of the particle  size and particle characteristics as well  as
depending on the field strength and free charge density.  Both  field
charging and diffusion charging occur in electrostatic precipitators with
diffusion charging dominating for particles less than  1  micron  in  radius.

Some  fairly refined  theories for  the  two charging mechanisms, field charg-
ing and diffusion charging, have been developed and  yield  results  which
are in reasonable agreement with experimental  data.   These theories and
available experimental evidence show  a minimum in  the  charging  rate for
particles in  the 0.1-0.5  urn range.  However,  the calculated  particle
migration velocities for  particles  in this range obtained  by assuming
reasonable  fields,  charge densities and  residence  times, as  shown in
Figure A-6,  are nevertheless quite  large.   It  would  seem that electro-
static devices  could be devised which could utilize  such large  particle
migration velocities for submicron particulate control.
                                    73

-------
4xlO-2
 u
 4
 V
 J!
 u
 •«
 i
                               Eo=2xl05V/m
                              No=1013ion/m3
                                t = 1 sec
                Diffusion^

            Field neglected
               0.1
0.2
0.3
0.4     0.5
                    Particle radius r_ , microns
   Figure A-6.  Calculated particle migration velocities
                        for submicron particles!/
                                 74

-------
Thermal Forces (Thermophoresis)

Particles can be removed from a gas stream by the use of a temperature
gradient.  The force which causes particle motion results from momentum
differences imparted to the particle on opposite sides.  The hotter  (and
thus faster) molecules colliding with the particle will impart a higher
momentum to the particle than the cooler (slower) molecules.  Aerosol
particles will then drift in the thermal gradient toward the cold sur-
face.  The motion of aerosol particles associated with a temperature
gradient is called thermophoresis.

                                                                      8/
Theoretical models for thermophoresis have been developed by Waldmann.-
Epstein,i/ Brock,-!^ Jacobsen and Brock!!' and Derjaquin and Bakanov..!=/
Thermal forces (i.e., radiometric forces), like other interactions
between gases and particles, depend on the Knudsen number (X/rp), where
X = mean free path of gas molecule and  rp = particle radius .

A number of equations have been developed to calculate the velocity  of a
particle due to the thermophoretic force.  All these equations can be
expressed in one general form:
                           v™ = - K -E- AT                      (A- 16)
                            T
 and differ  in  the definition  of  K  .  The factor  K  allows  for  the ef-
 fects  of  the thermal  conductivities of  the particulate material  and gas
 and includes various  accommodation  coefficients which are not  readily
 measured  or evaluated a  priori .  In the simple theories  K   is independent
 of particle size while in the more  complicated theories  K   also depends
 on particle size.   Table A-2  summarizes various equations that have been
 developed for  the thermophoretic velocity and the conditions for which
 they are  applicable.

 Figure A-7  presents the  results of  experimental measurements of  the  thermo-
 phoretic  velocity reported by Goldsmith and  May!!/ and Rosenblatt and

 La Mer.— '  The thermal  velocity is shown as a function  of — - —

 where   T  is  the mean absolute  temperature between hot and cold  plates
 and  dT/dX  is the  temperature  gradient across the plates.
                                    75

-------
              Table  A-2.   EQUATIONS FOR THERMOPHORETIC VELOCITY
Investigator
Wai dmflDzi

Epstein



Brock
Derjaniln and Bakanov
Thermo phore tic Velocity Equation
v_ s ~
T (8
vT=-l
2

VT = - 3
VT —
2
v,, = - I*
6P AT
+• TO)Tpg
u 1 kg
PgT 2k +
l_ e
u C1 1
PgT 1 + 6
kg + 0.5 kp


J*r
kp

r k* + °t ^ i AT
C^. V 1 + 2 kg/kp + 4 Cc X/dpJ
) " AT
Range of Applicability Comments
r « ) Shows good agreement with experimental
data
"' « rp Agrees with experimental measurements
for particles of low thermal con-
ductivity (k /k. £ 10)
?t « Tn Agrees with experimental results for both
good and poor conductors to within
= 25%
X « rn Agrees well with experimental results
(2 kg + kp)     pgT

-------
                40r
                30

             u

            I
             u
            m  20
            o
                10
                                              Goldsmith and  May
                                    Rosenblatt and  LaMer
0  20  40  60  80 100 120 140 160


  Temperature gradient


         CC/cm)
                                              dT_

                                          T   dx
Figure A-7.  The thermophoretic velocity in air as a function of tem-

                             perature gradient
                                  77

-------
Theoretical calculations as well as experimental measurements such as
those shown in Figure A-7 indicate that the thermophoretic velocity is
quite low and long residence times are required to accomplish significant
particle separation at reasonable temperature gradients.  Even at tem-
perature gradients of 10*°C/cm, the thermal velocity is only about 0.1
cm/sec (0.04 in/sec).  The large space requirements coupled with the high
cost of maintaining the required temperature gradient has resulted in no
industrial applications of thermal precipitators.

However, thermal forces might be exploited to enhance the performance of
certain types of control devices.  The deposition of particles from a hot
gas in a cooled, packed bed, is an instance  in which thermal forces may
improve collection efficiency.  Thring and Strauss report that experiments
have shown that when a packed bed is initially cold, particle collection
is more nearly complete.1^'  Strauss also estimated that, since the
passages in a bed are narrow, a temperature difference of 50°C might give
rise to a temperature gradient of 1000°C/cm in the passages.  Strauss's
calculations show that this would result in the deposition of 98.8% of
the particles of 0.1 micron diameter in a 9 in. deep bed.-±£'  Thring and
Strauss proposed to cool the packed bed by using it as an air preheater
in which another gas is heated as it cools the bed .is'

Diffusion Forces (Diffusiophoresis)

In a concentration gradient, which is accompanied by diffusion but not
necessarily by net motion of the gas phase, the heavier molecules will
impart a higher momentum than the lighter molecules.  If there is a net
motion of the gas phase (Stefan flow), additional force is applied to the
particles.  The combination of forces due to Stefan flow and the concentra-
tion gradient is referred to as the diffusiophoretic force.  Particle
movement by this force is called diffusiophoresis.

Several theoretical models of diffusiophoresis have been developed.
References 17-20 present some of the more pertinent theoretical work.
Equations (A-17) and (A-18) present expressions for the diffusiophoretic
velocity for the cases where  rp « X  and rp » X .
                                   78

-------
                                     D 7Yl    r
                                  n(mp-mv  AC   r     ,         /A in\
                           vp = D - *- - —   rp » *•         (A- 18)
                      (1+rra
where u A^ =
          a^ = coefficient of thermal reflection,

           n = total number of gaseous molecules per unit volume,

          — = vapor concentration gradient,
          AX
           0 = density of gas vapor mixture,

          m  = particle mass,

          tiiy = vapor molecular mass, and

           D = diffusion coefficient.

Deryagin reported experimental results which show  that Eq.  (A- 18) holds
for Knudsen numbers  < 0.5  and  that Eq.  (A-17) holds for Knudsen numbers
>  0.7.—

Figure A-8  presents experimental measurements of  the diffusiophoretic
velocity in helium  (47. 02) reported by Goldsmith  and May.i2/  In  the
experiments of Goldsmith and May, a saturated salt  solution was used on
one wall of the  vapor pressure  gradient box with  water  on  the other wall.

Diffusiophoresis should be a useful mechanism to  exploit in conjunction
with  other  mechanisms for the removal of  small  (< 2 microns) particulates
from  gas streams.  Its use has  the following advantages:

1. The  fundamental mechanism is  independent of particle size and becomes
more  important compared  to other  removal  mechanisms for particles below
2  microns.
                                    79

-------
                0.12
20-25° C
25-30° C
30-35° C
35-40° C
40-45° C
45-50° C
                    0  10  20  30 40 50  60  70
                    Water vapor gradient mg/cm
                                                       v
                                                       o
                                                       •
Figure A-8.  The diffusiophoretlc velocity in helium as a function of
               water-vapor pressure gradient at temperatures between
               20° and 50°C
                                  80

-------
2.  The particulate removal efficiency that can be expected depends on
operating conditions and equipment design but can theoretically reach
100%.

To exploit the mechanism of diffusiophoresis, the collecting device must
be designed such that one component in the gas phase is diffusing toward
a collecting surface.  The most practical case is the diffusion of water
vapor toward a surface.  Recent studies by Calvert, et al.£i' have shown
that diffusiophoresis can produce appreciable particle deposition under
the heat and mass transfer conditions of a realistic wet scrubber.

Magnetic Forces

A force (Lorentz force) is generated when an electrical charge or an elec-
trically charged particle moves in a magnetic field transverse to the
field lines.  If a dust particle carrying "n" elementary charges "q" moves
with a speed  v  , the direction of the force will be at right angles to
both the direction of the field and the direction of motion of the par-
ticle so that the particle will be diverted from its original path.  As
a result of the change in direction of the particle, the possibility of
particle precipitation exists.

The potential for particle precipitation using the Lorentz force can be
assessed by determining the terminal drift velocity of a particle in a
magnetic field.  The terminal velocity can be obtained by equating the
Lorentz force to the resistance of the gas, calculated from the Stokes-
Cunningham law:
                           nqvB = SiruDpVmC                     (A- 19)

where   n = number of charges on particle,

        q = elementary charge,

        v = velocity of particle in field,

        B = magnetic field strength,

        u = gas viscosity,

       Dp = particle diameter,

       vm = terminal drift velocity of particle, and

        C = Cunningham factor.

                                   81

-------
The left-hand term in Eq. (A-19) represents the Lorentz force and the right
term the gas resistance.  Equation (A-19) can be rearranged to yield the
following expression for the terminal drift velocity:
                                                               ,. or.N
                                                               (A'20)
Equation (A-20) indicates that the terminal drift velocity varies directly
with the number of charges on a particle, the particle velocity, and the
magnetic field strength and inversely with the particle diameter.

Examination of Eq. (A-20) indicates that the terminal drift velocity is
less than 0.01 cm/sec for particles 0.1-1.0 micron in size except at high
particle velocities and magnetic field strengths.  The above theoretical
considerations demonstrate that the force generated by a charged particle
moving in a magnetic field is too small for use in particle separation.
Furthermore, because of the low values of the terminal drift velocity at
more nominal conditions of particle velocity and magnetic field strengths,
the use of magnetic fields to agglomerate particles does not appear
feasible.

PARTICLE AGGLOMERATION AND /OR PARTICLE GROWTH

The agglomeration or growth of fine particulates , as a step in the con-
trol of fine particulates, is in principle attractive.  If large par-
ticles can be produced for reasonable energy expenditures, it may be
possible to utilize devices which cause particle agglomeration or growth
in conjunction with conventional control systems.  Condensation is the
most feasible way to achieve particle growth without agglomeration.  Con-
densation phenomena are discussed in detail in Appendix C, and only par-
ticle agglomeration will be discussed in this appendix.

The movement of particles toward one another may be brought about solely
by Brownian motion, but usually other influences play a major role, e.g.,
turbulence of the fluid, gravitational forces, and electrostatic forces.
The following sections discuss the nature of particle agglomeration with
and without the use of external forces.

Agglomeration in Absence of External Forces

When the movement of particles, leading to contact and agglomeration, can
be effected only by Brownian movement (diffusion), the process is called
                                   82

-------
thermal coagulation.  For the general case of a polydisperse  aerosol,  the
number of particles of radius  r2  diffusing in unit  time  to  a  fixed par-
ticle with radius  r^  is given by—'
                  N12 = 4n  (D!+D2)  (ri+r2) n  (r2,t)             (

By expressing the diffusion coefficients in terms of  the particle  mobility
0 (0 - D/kT) and introducing the coagulation  constant  KQ  , Eq.  (A-21)
can be written as
                      N12 = K0 (r1>r2) n  (r2,t)                 (A-22)

where  KQ (r^) = 4rr (0!+02) (ri+r2) kT   .
The basic equation for the change with time of the size distribution of
aerosols due to thermal coagulation is given by Muller.ZJi/

                   ml 2
         dn  (m
 t)    r
*—*- = /  K (m.,m-m) n (m.,t) n (m-m-,t) dm.
                 00
         n (m,t) /K (m,mi
            at
                  0            oo
                                    i,m]_)  n (m]_,t)  dm]_,
where  n (m,t) dm  is the number of aerosol particles per cubic centimeter
at the time  t , whose masses lie between the values  m  and  m + dm  ,
while  K (m^,m2) gives the value of the coagulation constant between
aerosol particles of the masses  m^ and n^ .  The physical meaning of
this integro-differential equation is that a particle of mass  m  can only
come into existence if two particles with masses  m,  and (m-m.) collide:
it can be integrated either from  m, = 0 to m, - m/2  or from  m^ - 0 to
m, = m .  The integral value must be halved as, otherwise, each possible
combination is counted twice.  The second integral expresses that each
particle of mass  m , disappears from the fraction  m to m + dm  after
colliding with a particle of mass  m^ .

By making simpler assumptions Smoluchowski ^t*-^' obtained, instead of
the integro-differential equation, a simple differential equation which
is easily solved.  If only the concentration of the total number of
                                   83

-------
particles, n , is considered and the same value of  Ko , the coagulation
constant, is used for all particles, then the fundamental equation of
coagulation becomes
                                  £°. n2  .                     (A-24)
The factor  1/2  is again necessary so that each aggregate should be
counted only once.  The solution of Eq. (A-24) is
                       -   _ =  > t or n --  -          (A.25)
                     n   no   2           1 +
where  no  is the initial particle number concentration at time  t = 0.

Equation (A-25) for the particle number concentration as function of
time, t , is a good approximation for the size distributions occurring
in many aerosols.

The major disadvantage of thermal agglomeration with regard to partic-
ulate collection is the long time required to grow particles.  The only
variables which can be used to control particle-particle collision rate
are temperature, gas composition and the particle concentration.  The
temperature and gas composition do not appear to be useful variables.

The possibility of influencing the coagulation rate of particles suspended
in air by introducing a second gas or vapor has been explored in several
experimental programs, and the only clear-cut results seem to be those in
which the added vapor in some ways affects the shape of the particles,
or in which vapor is actually being transferred between the vapor phase
and the particles so that appreciable gaseous diffusion occurs  (i.e.,
diffusiophoresis) .

One could conceivably change the particle distribution and concentration
by seeding the particle suspension with  large particulates to act as
"agglomeration sites," but this does not appear very useful considering
the  n2  dependence of the total agglomeration rate.  Thus, thermal
agglomeration  is not a viable approach to augmenting the ability of con-
trol systems to control fine particulate emissions.
                                    84

-------
Turbulent Agglomeration

Turbulence increases the relative velocities among participates which in
turn increases the chance of particulate collision.  Theoretical studies
on the coagulation of aerosols in turbulent flow have been conducted by
Levich ,!§/ East and Marshal I,2-!/ Tunitskii.i^/ Obukhov and Yaglon,!2' and
Levich discusses coagulation effected by fluctuations having a scale of
the same order as the particle size, which is appreciably less than the
internal scale of tubulence  \o .  Levich derived the following equation
for the coagulation rate
1/2
                       = 32TTrVo « 25  l      r3 nQ           (A-26)
where   p = coagulation constant,

        Y = kinematic viscosity, and

        e = rate of dissipation of turbulent energy per gram of

              medium
The ratio of the constants of turbulent and thermal coagulation for  this
case is
                                / \ -\ n  s  x
                                                               (A-27)
where  KI   is a constant of  the order 0.1-0.5.  For particles   1 um  in
radius   r2/ ra 0.1  .  Hence  it follows  that for the coagulation of  such
         p/D                                          i	
particles  to be accelerated  noticeably  by  turbulenceis/e/v  must be  about
100 and  ^o about 0.1 cm.  This in  turn requires a very high velocity.
Under these conditions, the  effect  of turbulent coagulation is  negligible
for particles of radius 0.1  urn while for particles of 10 um it  is very
large.

Beal has studied the case where the sink particle is  larger than  the
turbulent  microscale (rp> X0) and has derived  the following equation for
the collision rate  per unit  area of sink particle—'
                                    85

-------
                                                               CA-28)
where  P^ = constant, ~ 1/4 .

For this case the ratio of constants of turbulent and Brownian diffusion
coagulation is
Figures A-9 and A-10 present some of the results of Real's theoretical
calculations.  As shown in Figure A-9, the "exact" solution of the dif-
fusion equation is asymptotic to the solutions based on either Brownian
motion or turbulent diffusion alone and does not differ very much from
a simple sum of these solutions.  Figure A-10 presents the normalized
coagulation constant, K , as a function of particle diameter.

Seal's analysis has several limitations.  Only the steady-state solution
of the diffusion equation was obtained, only interactions among particles
of equal size were considered, and all particles were assumed to stick
together on impact.  The last assumption is the most tenuous.

Experimental study of the rate of agglomeration in the presence of tur-
bulence is difficult because turbulence also accelerates the deposition
of particles on walls.  The rate of deposition increases with particle
size so that the assessment of the course of agglomeration is complicated.
Experiments by Yoder and Silverman are the only results that appear to be
subject to any rigorous analysis.—'  These investigators performed ex-
periments to obtain data on the deposition and agglomeration of particles
in turbulent air flow.  In their experimental design, deposition and
agglomeration were occurring simultaneously, and their basic problem was
to separate the two effects.  They did this by measuring both the total
number concentration and the fraction of particles which had agglomerated
at the inlet and outlet of their test section.  By applying certain
theoretical concepts, they could then infer the separate effects of both
deposition and agglomeration from the measured parameters.  They did not,
however, make any direct observation of either phenomenon.
                                   86

-------
      l\llllll|	1  I I
            FLUID AIR
            PIPE PIAMTER 10 CN
            TEMPERATURE !0*C
 KT«  t
              t    •   »»  I   B  10'  t
                PMTICIC DUUKTCR. fin
Figure A-9.  Particle agglomeration  flux
   for various  particle diameters and
   fluid velocities
       FLUID-AIR.
       TEMPERATURE* ZO'C
       PIPE OIAMETER'IO CM
       PARTICLE DENSITY* 1.0 GM/CC
          Q«3,000 CM/SEC
          3-1.000 CM/SEC
            300 CM/SEC
          u-100 CM/SEC
            I I Hill    I  I  I I mil
10*
    Iff* t
            5   10*'  S    S   10°  t
              PARTICLE DIAMETER,
   Figure A-10.   Coagulation constants
     for various  particle diameters
     and fluid velocities
                     87

-------
Beal utilized  the data of Yoder and Silverman for a comparison with his
theoretical predictions of particle agglomeration.  Figure  A-ll presents
this comparison.  The  agreement is reasonably good for  the  0.8 urn par-
ticles, but poor for the 0.26 um particles.  No adequate explanation of
the difference in magnitude between the observed and predicted values
of the coagulation  constant was offered by Beal.  A general discussion
of the difficulties encountered in the experimental determination of
turbulent agglomeration coefficients is presented by Fuchs.^i'
                          to-*

                          9
    1  1  l l i i ill
                                O 0.26/iml VODER AND
                                O 080/imJSILV£RMAN'S D4TA
                               	 PREDICTED
                                                 I 11II
K>»  2    5   K>«  2    S
       REYNOLDS NUMBER
                                                    ios
     Figure A-ll.   Comparison of theoretical predictions with  Yoder
                          and Silverman1s agglomeration data.
                                    88

-------
Figures A-9 and A-10 indicate that turbulence does not significantly en-
hance the coagulation constants for particles less than 0.2 urn in diameter,
but can increase the coagulation constant of 0.5 um particles by about  a
factor of 10 and 1 um particles by a factor of 102.  The energy expended

to accomplish this increase is about 2.7 x 107 dyne-cm-sec"  Qr Oi004
                                                    g
hp/g.  This is equivalent to about 40 hp/1,000 ft3.  If this energy is
expended over a period of 1 min, a 0.5 um particle might be grown to about
2-3 um.  This is an extremely high energy consumption and, therefore, tur-
bulent agglomeration does not appear attractive if an external source is
required to provide the turbulence.
Agglomeration in Sonic Fields

A potentially effective way of rapidly agglomerating fine particles  to a
larger size is to pass the aerosol through a sonic field.  Several dif-
ferent effects are responsible for the enhanced agglomeration rate due
to sonic forces including  (1) collection of the particles at the antinodes
in the sonic standing wave (due to radiation pressure),  (2) hydrodynamic
forces between the particles and, (3) additional collisions due to the
different vibrational amplitudes of different sized particles.  Although
no comprehensive theory is available some idea of the parametric dependence
of these several forces can be obtained from standard equations.

The radiation pressure exerts a maximum force  F  on a small, rigid  sphere
in a frictionless gas which is given by the expression!?./
                        F =  5/3TT2  pe IfSi   r3                   (A-30)
                                    6   \s     V


where   ^s   is  the wavelength of  the sonic vibration  pg   is  the  gas density,
and  Av is velocity amplitude of  the  gas relative  to the particles.

The hydrodynamic potential between two particles  i|>(d)   is given by


                                  (sTn2 0.2  c5s2 9)              (
where   9   is  the  angular  difference  between the direction of vibration
and  the line  between the  centers  of  the particles  and  d  is the dis-
tance between the particles.
                                   89

-------
The vibrational amplitudes of different size particles can be obtained
from the expression of Hiedemannl^'
                       X
                        g
1/2
                                                               (A-32)
where  Xp  and  Xg  are the vibrational amplitudes of the particle and
the suspending gas and  f  is the frequency of vibration.

The relative importance of the three mechanisms cited above has not been
determined.  In fact, there is no assurance that other mechanisms are
not important.  It is therefore not possible to construct a theoretical
model even if the equations given above could be applied to real situations,
Nevertheless, it is interesting to note the functional dependence of the
equations for these mechanisms.  All three expressions show a strong de-
pendence on the radii of the particles, with the forces due to radiation
pressure and hydrodynamics increasing with increasing particle size and
the relative vibrational amplitudes decreasing with increasing particle
size.  There is also a strong frequency dependence in all these expres-
sions.  Since the value of  Au  drops to zero at low frequency, it can
readily be seen that all three mechanisms for particulate agglomeration
will become unimportant at low frequency.  At high frequencies, on the
other hand, the vibrational amplitude of the particles become negligible.
There is therefore an optimum frequency for acoustic agglomeration which
varies with particle size.

As noted in the previous paragraphs there is no usable theoretical model
available for sonic agglomeration processes.  There is, however, a large
body of rather diverse experimental results available and Ref. 35 presents
an extensive review.  Several generalizations can be made from this body
of literature.  Since an adequate model does not exist, these generaliza-
tions, which are listed below, provide a basis for consideration of pos-
sible practical application of sonic agglomeration.

1.  Sonic fields can effectively induce agglomeration in all sizes of
suspended particles, including those in the submicron range.

2.  All materials can be agglomerated.

3.  The  efficiency of the agglomeration depends upon the square of the
particulate concentration.  The minimum particulate loading that can be
treated  is usually given as 1.1-2.3 g/m3  (0.5-1.0  grain/ft3), although
number density  and mean size would appear  to be more important parameters.
                                    90

-------
4.  The sonic agglomeration process is strongly frequency dependent,
with the optimum frequency varying with the particle size or sizes  to be
agglomerated.  Optimum frequencies range from a few hundred to an upper
limit of about 10,000 Hz.

5.  The required residence time for a given degree of agglomeration
varies with field intensity, and minimum field intensities of about
140 dB are required for usable residence times.  For a 150-160 dB field
a residence time of 5 sec is typical.

A number of industrial installations of sonic agglomeration systems are
discussed in Ref. 35.

Sonic coagulation of participates has as its principal advantage its
applicability to any aerosol, including those comprised of submicron
particles.  The principal disadvantage of sonic coagulation is its  rela-
tively high energy requirements, which have been reported ranging from
0.05-0.7 hp/m3/min (1.5-19 hp/1,000 cfm).  It is this factor which  has
apparently led to the virtual abandonment of sonic precipitators.  How-
ever, Mednikov3!' suggests that minimum total energy requirements of
0.16-0.32 hp/m3/min (4.5-9 hp/1,000 cfm) should be achievable with  optimum
design.  Presumably this energy requirement is based on agglomeration
to a sufficient size to be removed by inertial precipitators.  A second
major disadvantage is the low efficiency of acoustic coagulators and their
inability to handle highly dispersed suspensions.  Even with long resi-
dence times sonic precipitators which incorporate inertial separators
cannot treat suspensions having particle loadings of less than 1.2-2.3
g/m3 (0.5-1.0 grains/ft3), and leave a somewhat lesser quantity of  par-
ticles in suspension.  It is therefore necessary to augment highly  dis-
persed suspensions with a water mist or other particles to increase the
particle loading.

From the above discussion, the energy requirements of a sonic precipitator
with an inertial separator or other simple filtering system seem rela-
tively high.  Water augmentation appears necessary to obtain the removal
of a large fraction of the particulates.  If 0.02 kg of water per cubic
meter per minute is used, one would expect an increase in the power re-
quirements of up to 0.04 hp/m3/min, depending on the device used and the
mean droplet size.  The minimum energy requirements would then be 0.2-
0.35 hp/m3/min (5.5-10 hp/1,000 cfm).  Whether this is competitive with
other devices capable of removing fine particles remains to be seen.
                                   91

-------
Alternatively, one should consider hybrid units.  For example, sonic
agglomeration could be used upstream from a conventional electrostatic
precipitator.  Since the particulates now need to be agglomerated to
only 1-2 urn instead of 10-15 urn with an inertial separator, a reduction
in acoustic energy requirements by at least a factor of 5 would be ex-
pected.  Other hybrid units should also be possible.

Agglomeration of Charged Particles

One method of increasing the rate of agglomeration of fine particulates
is to add a bipolar charge, either with or without an externally imposed
field.  With proper conditions large electrostatic forces between par-
ticulates can produce a large increase in the rate of agglomeration of
submicron particulates.

Fuchs!2/ and Zebell^/ have derived the following equation for the rate
of agglomeration for charged particles:
                           - S2 = k  K n2                      (A-33)
                             dt    em o

where   KQ = agglomeration coefficient in the absence of charge, and

       kem = correction factor to allow for both particle charge and
               particle mean free path.
The correction factor, k^  is given by

                                     N
                                      -9	                    (A-34)
where   N
                                    92

-------
         k = Boltzmann's constant,
   D   , D   = particle diameters,
    Pl   P2
     m,, 102 ~ particle masses,

   q    q   _ charge on particles in electrons,
    Pl'  P2
         eQ = permittivity of free space,

         ef = dielectric constant of gas medium,

          T = temperature, and

     Dl» D2 = particle diffusion coefficients.
Nq  is a parameter to allow for charge effects and  Nm  is a parameter to
allow for slip between particles (as distinguished from slip between
molecules) in the collision between two particles of different diameters
D   and D   , diffusion coefficients  D^ and D2  and masses  m^ and nu .
 p!      P2
Figures A-12, A-13, and A-14 illustrate the Influence of charged particles
on agglomeration for a monodisperse aerosol in which all particles have
the same magnitude of charge.1Z/  Figure A-12 presents the agglomeration
rate in terms of number concentrations while Figure A-13 shows it in
terms of mass concentration.  The effect of particle charge on agglomera-
tion rate is shown in Figure A-14.  this figure shows that a unipolar
charge causes a very rapid reduction in  kem , and hence in agglomeration
rate, while the rate of increase in agglomeration rate due to ambipolar
charge is not as rapid.  Figure A-14 shows only the relative effect of
charge on agglomeration rate for various charge levels and particle sizes.
It implies a very large relative effect for large particles but tells
nothing about the absolute rate directly, since the particle agglomera-
tion rate decreases rapidly with increasing particle size for a given
mass concentration.  The absolute effect can be obtained by considering
Figures A-13 and A-14 simultaneously.

For weakly charged monodisperse aerosols, Figure A-14 indicates that there
will be little change in the agglomeration rate for fine particles.  For
highly charged aerosols, the agglomeration rates can be much higher (up
to 100-1,000 times that of the uncharged aerosols), but will usually fall
                                    93

-------
               co
                E
               -21
                o
                   10°
             Q O
             in —
[•dnp
npdt

K°kemn
Ko = 2.98xlO-10cm3/sec
                                         FOR fj. = 1.8xl(T4 poises

                                              T=25°C
U    ID"4     ID'3    ID"2      10-1
                                                        1
                                              10
                     AGGLOMERATION RATE,
                                             npdt
                                , percent particles/second
 Figure A-12.  Effect of  particle number concentration on agglomeration rate!!/
           O

           1
           Z
           O
           U
           to

               1,000,000
100,000
                  10,000

           Q
           i
                    1,000
                     100
                      10
              p • 1 grn/cm3

              T=25°C
                              .„..)
                                             ii^uL
        1
-------
        1000
VO
Ui
                    SPECIFIC PARTICLE
                    SURFACE GRADIENT,
                       , Volts/Micron:    1
                                                                                            4kT
                                                                                    E = Specific Particle Surface
                                                                                                    100
                                           PARTICLE DIAMETER. Dp, Microns
                      Figure A-14.   Effect of particle charge on agglomeration—!

-------
rapidly due to the annihilation of the charge upon agglomeration and de-
crease in number density.  This result does not seem very promising for
practical applications.

However, if we consider a bipolarly charged polydisperse aerosol contain-
ing a large number of relatively large particulates as well as the fine
particulates the situation appears much more favorable because it may be
possible to collect the fine particulates on the large particulates.  As
a result fewer steps would be required to substantially reduce the fine
particulate concentration.  Furthermore, the agglomeration process would
result in much less charge annihilation.
                                   96

-------
                             REFERENCES

 1.   Wong,  J.  B.,  W.  E.  Ranz, and  H.  F. Johnstone, J. Appl. Phy... 2(6,
       244-249 (1955).

 2.   Ranz,  W.  E.,  Contact Report AT (30-3)-28; University of Illinois
       Technical Report  No.  3 (1951).

 3.   Friedlander,  S.  K., A.I.Ch.E. Journal.  ,3, 43  (1957).

 4.   Langmuir, I., U.S.  Office of  Scientific Research and Development
       No.  865, Part  4,  October 1942.

 5.   Ranz,  W.  E.,  and J. B.  Wong,  Ind. Ene.  Chem.. 44,  1371-1381  (1952).

 6..   Kraemer,  H. F.,  and H.  F. Johnstone,  Ind. Eng.  Chem.. 47, 2426-2436
       (1955).

 7.   Robinson, M., "Electrostatic  Precipitation," Air Pollution Control.
       Part I. Werner Strauss, Ed., Wiley—Interscience, New York (1971),

 8.   Waldmann, L., Z. Naturforsch.. 14a, 489 (1959).

 9.   Epstein,  P. S.,  Z.  Physik. 54, 537  (1929).

10.   Brock, J. R., "On  the Theory  of  Thermal Forces  Acting on Aerosol
       Particles," J. Coll.  Sci..  17, 768-780  (1962).

11.   Jacobsen, S., and  J. R. Brock, J. Coll. Sci.. £0,  544  (1965).

12.   Derjaguin, B. V.,  and S. P. Bakanov,  Dokl. Akad. Nauk.. SSSP (phys.
       Chem.,  USSR),  147,. 139 (1967).

13.   Goldsmith, P., and F. G. May, "Diffusiophoresis and Thermophoresis
       in Water Vapor Systems," Aerosol Science. C.  N.  Davies, Ed.,
       New York, Academic Press, 164-194  (1966).
                                   97

-------
14.  Rosenblatt, P.,  and V.  K.  LaMer,  Phys.  Rev..  70,  387  (1946).

15.  Strauss, W., and M. W.  Hiring,  "Studies in High Temperature Gas
       Collecting," Journal  of  the Iron & Steel Institute.  196, 62  (1960),

16.  Strauss, W., Industrial Gas Cleaning. Pergamon Press,  Oxford,
       215-231 (1966).

17.  Schmitt, K. H.,  and L.  Waldmann,  Z. Naturf..  15 a, 843  (1960).

18.  Deryagin, B. V., and S. S. Duklin, Dokl. Akad. Nauk.,  SSSR (Phys.
       Chem.), 106, 851; and   11, 613 (1956).
19.  Sparks, L. E., and M. J. Pilat,  Atmos.  Environ..  4,  651  (1970).

20.  Deryagin, B. V., and S. P. Bakanov,  Dokl.  Akad. Nauk.. SSSR (Phys.
       Chem.), 147, 139 (1962).

21.  Calvert, S., et al., "Feasibility of Flux  Force/Condensation
       Scrubbing for Fine Particulate Collection,"  EPA Report No. EOA-
       650/2-73-036, October 1973.

22.  Muller, H., Kollordzentschrlft.  38,  1 (1926).

23.  Muller, H., Kollordzentschrift.  £7,  223,  1928b.

24.  Smoluchowski, M. von., Phys. Z.. 17_, 557-585 (1916).

25.  Smoluchowski, M. von., Z. Phys.  Chem..  _92, 129 (1818).

26.  Levich, W., Dokl. Akad. Nauk.. SSSR, 99,  809,  1954a.

27.  East, T. W. R., and J. S. Marshall,  Q.  Journal R. Met. Soc.. 80,
       26 (1954).

28.  Tunitsky, N. N., Zh. Fiz. Khim. , 20-, 1136 (1946).

29.  Obukhow, A., and A. Yaglom, Prikl. Mat. Mekh.. 1.5, 1 (1951).

30.  Seal, S. K., "Turbulent Agglomeration of Suspensions," Aerosol
       Science. 3, 113-125  (1972).
                                   98

-------
31.  Yoder, J. D., and L. Silvennan, "Influence of Turbulence on Aerosol
       Agglomeration and Deposition in a Pipe," Paper No.  67-33, 60th
       Annual Air Pollution Control Association Meeting,  Cleveland,  Ohio,
       13 June 1967.

32.  Fuchs, N. A., The Mechanics of Aerosols.  Pergamon Press, New York
       (1964).

33.  King, L. V., Proc. Royal Soc.. A147, 212  (1934).

34.  Hiedemann, E., Kolloidzeitschrift. 77, 168 (1936).

35.  Mednikov, E. P., Acoustic Coagulation and Precipitation of Aerosols.
       Consultant Bureau, New York (1965).

36.  Zebel, G., "Coagulation.of Aerosols," in Aerosol Science. C. N.
       Davies, Ed., Academic Press, New York (1966).

37.  Lapple, C. E., '"Electrostatic Phenomena With Particulates," in
       Advances in Chemical Engineering. Vol.  8, Academic Press.
                                   99

-------
                           APPENDIX B

     EMERGING CONTROL TECHNOLOGY—SPECIFIC CONTROL DEVICES

INTRODUCTION

A variety of collection devices touted as being effective for the col-
lection of fine particulates, have been reported in the literature or
submitted to EPA for evaluation.  The devices range from laboratory-scale
devices to full-scale units that have been tested on industrial sources.
Several of these devices were evaluated during this study and specific
systems are reviewed in the following sections of this appendix.
                                    100

-------
Steam-Hydro Air Cleaning System
Lone Star Steel
Lone Star, Texas

Lone Star Steel has recently developed a wet scrubbing system which
utilizes a high-speed steam drive with injected water to accomplish the
removal of particulate pollutants.  The system consists of a steam noz-
zle, water injector, mixing tube, and twin cyclones.  Normally, the
system operates on energy produced by waste heat captured from the pro-
cess being controlled.  The heat is used to generate steam in a waste
heat boiler.  In installations where heat energy is low, supplemental
heat may be provided.  In many cases, a package steam boiler may supply
all the energy.  In addition to driving the system, the steam nozzle
creates a draft which draws contaminated gases into the system.

Figure B-l presents a simplified diagram of the system.!.'   The first
stage of cleaning is done in an optional atomizing chamber with water
sprays that may be employed to cool the gas stream and remove heavy
particulate.  The conditioning or atomizer chamber is usually only neces-
sary if the gas stream contains submicron particulates.

The steam nozzle, a proprietary design, surrounded by eight spray nozzles
as shown in Figure B-2 is a key element in this system.!'   The exit
velocity of the steam is about 1,050 m/sec (3,500 ft/sec).  The function
of the steam is to breakup or shear the surrounding water spray and to
provide motive power for moving the gas stream through the system.  The
flow rate of injected water controls the cleaning efficiency.  The steam
flow rate has little to do with the cleaning efficiency, but some is re-
quired to shear the injected water spray.

Steam requirements as reported by Lone Star Steel range from 0.03-0.16 kg
of steam per cubic meter  of gas (2-10 lb/1,000 ft3) depending on up-
stream process draft requirements.  Water rates are 0.23 kg of H20/kg of
gas (230 Ib of H£0/1,000 Ib of gas) for,, the conditioning or atomizing
chamber and for the spray nozzles surrounding the steam jet.

Efficiency tests have recently been performed by an EPA contractor* on
a unit installed on multiple open-hearth (oxygen lanced) steel furnaces
at Lone Star Steel.  Figure B-3 presents a diagram of one of the units.
   Tests performed by Southern Research Institute using impactors and a
     diffusion battery for particle sizing and fractional efficiency
     measurements.
                                   101

-------
                                                            Cyclones
Figure B-l.  Simplified diagram of Steam-Hydro Air Cleaning System
                                 102

-------
        Steam
s
                                     Water Spray
                                     Nozzles (8)
Figure B-2.  Diagram of steam nozzle and water spray nozzles (Steam-
                           Hydro Air Cleaning System)
                                     103

-------
                          «250lb/min
                               H2O

                                 25 psig
                 Steam -
                 250 psig
                 (Satrd)
                 150lb/min
              Mixing Tube
               18"dia.xlO' Lg.
            Location of —
Standing Shock

   Impaction
               Section
                 w2501b/min
 Dirty Gas From
 Waste Heat Boilers
 1000-1200lb/min
  500-600 °F
 Avg  1.3 Grains/SCF
 140°F
          jpoos
                                                               Clean  Gas to
                                                               Atmos.(Satrd)
                                                                w.OOl Grains/SCF

                                                                 t
                          Droplet
                          Growth
                          Sect-ion
 +0.5"W.C.

+2"W.C.
 /\  /\ A
                              /\
 A  f\ A

-7"W.C.
         -7"W.C.
                                          +2"W.C.
                                          + 10"W.C.
                           Conditioning
                           Chamber
                        (6"diaxlO'High)
                        48 "dia
                              Dual Cyclones
                              Operating in  Parallel
                              (AP = 2"W.C.)
Figure B-3.   Diagram of one Steam-Hydro air cleaning unit serving open-hearth furnace

-------
Preliminary test results indicate collection efficiencies of 99.57. or
greater.  During the tests, inlet grain loadings averaged 2.88 g/m3
(1.26 grains/scf), but varied from 0.23-6.86 g/m3 (0.1-3.0 grains/scf)
depending on the point in the process cycle.  Outlet grain loadings
ranged from 0.002-0.02 g/m3/(0.0008-0.0089 grains/scf).  The inlet
size distribution of particles had a mass mean diameter of 1.5 microns
and nearly all the particulates were less than 5 microns in diameter.

Capital, operating, and maintenance cost data are not available for the
device.  Lone Star Steel has stated that "the installed cost is more
than competitive with conventional devices (excluding cost of waste heat
boilers)."!/

Based on the results of the preliminary tests conducted by SRI, the Steam-
Hydro system offers excellent promise as a new system for the control of
fine particulate emissions.  Hie system appears to be applicable to a
wide range of sources, but primary-applications are likely to be sources
where waste heat is available and which emit fine particulates.  Where
waste heat is not sufficient to power the system, supplemental heat may
be used.  However, the use of supplemental heat would add to the overall
cost of the system.
                                   105

-------
ADTEC Wet Scrubbing System
Aronetics, Inc.
Tullahoma, Tennessee

The ADETC system, developed by Aronetics,  Inc., Tullahoma, Tennessee,
is a wet scrubbing system that operates on the conventional venturl col-
lection-mechanism of inertial impaction, but establishes the requisite
particle-droplet differential velocity by utilizing waste process heat
rather than external energy.  The Aronetics  system is shown schemati-
cally in Figure B-4.1'  Waste gas from the industrial process passes
through the heat exchanger, the two-phase jet, and the separator, in that
order.  Water  is pumped at 300 psi to a heat exchanger where heat is
transferred to the water from the gas increasing the temperature of the
water to the required level, usually between 300°F and 400°F.  Expanding
through the nozzle of the two-phase jet, partial flashing occurs (approxi-
mately 15% vaporizes), and the remaining liquid is atomized  (water drop-
lets average approximately 10 urn).  Thus, a two-phase mixture of steam
and small water droplets leaves the nozzle at high velocity.  The dirty
gas is entrained by this high velocity two-phase mixture, and in the
ensuing mixing with the gas, cleaning occurs primarily by impaction.  At
the same time, transfer of momentum to the gas results in a pressure rise
across the mixing section.  This pressure rise provides the motive force
to move the gas through the system.

The mixture of steam, gas, and entrained water with dust collected in
cleaning, passes through the separator after discharging from the mixing
section.  The water with the dust is removed from the gas and steam mix-
ture which then leaves the separator through the stack.  The water with
dust is discharged from the separator to the wastewater treatment sub-
system.  Chemicals and other treatment are applied to remove the dust
and other contaminants from the water for recovery of the water and to
prevent scale  and corrosion in the system.

The treated water is returned to the pump for recycling and  sludge is
removed for disposal by other pumps not shown.  Make-up water to replace
water leaving  the system through the stack in vapor form and with the
sludge is introduced in the separator where some heat is transferred to
the make-up water from the gas.

Since the transfer of momentum produces a pressure rise across the mix-
ing section which provides the motive force to move the gas  through the
system, the ADTEC system does not require, nor does it include, a blower
which is necessary in other cleaning systems or methods.  A  typical
                                   106

-------
                                                        •STACK
         TWO-PHASE
         JET NOZZLE
  HEAT EXCHANGER
•MIXING
 SECTION
                          PUMP
                                                       •SEPARATOR

                                                        -MAKE-UP
                                                         WATER
                                         WASTE WATER
                                          TREATMENT
                                              t
Figure B-4.  Schematic diagram of basic Aronetics System
                           107

-------
pressure rise plot as a function of water temperature is shown in Figure
B-5 for a water-to-gas ratio of 1-kg water per kilogram gas.  System
pressure loss in some cases may include 35.6 cm (14 in.) water gauge in
the ducting and heat exchanger, and 20.3 cm (8 in.) water gauge in the
separator for a total of 56 cm (22 in.) water gauge.  From Figure B-5,
a water temperature of 360°F would provide the necessary pressure rise
of 56 cm (22 in.) water gauge.

Water temperature and water-to-gas mass ratio are the two principal de-
sign parameters.  Values for each are selected in a specific design to
meet cleaning and pumping requirements for each application.  In general,
fine dust will require higher temperatures and higher water-to-gas ratios.

A test unit was installed on a ferroalloy furnace of the Chromium Mining
and Smelting Corporation (Chromasco) at Woodstock, Tennessee.  Test re-
sults from the pilot system are given in Table B-l.3-/  From left to right,
the columns give the test number, the furnace product, the system con-
figuration, the water-to-gas ratio, the water temperature, and the outlet
dust loading.  Several variations in equipment configuration were made
during the test program.  These variations are noted by a letter designa-
tion in the Aronetics system configuration column.  Inlet dust loadings
ranged from a low value of 0.62 g/m3 to a high value of 19.8 g/m3.  There
was no apparent effect on outlet dust loading as a result of increased
inlet loading.  Collection efficiencies of 99+% were reported.  The sam-
pling and analysis methods used to obtain the data shown in Table B-l
are not known.

Particle size distributions of the dust to which the results of Table B-l
apply are given in Table B-2.  Furnace products during certain tests were
75% and 50% ferrosilicon.  Size distributions in the reaction gas differ
slightly for the two products, and both are given in Table B-2.  Although
all of the dust is finer than 0.5 urn, the most significant aspect of
these data are the large amount, 80% plus by weight, which is < 0.1 urn.
The size distributions reported in Table B-2 were obtained by optical pro-
cedures using an electron microscope.i'

Mechanical energy consumed in pumping the water to the required pressure
of 275 to 300 psi is about 0.11 hp/m3/min (3 hp/1,000 cfm) of gas while
the power required in the form of heat input to the water from the gas
stream exceeds 7.0 hp/m3/min (200 hp/1,000 cfm).  The latter power re-
quirement would obviously be excessive if it were not supplied by waste
heat from the gas stream.  Capital costs have been estimated by Aronetics
to be between those of venturi scrubbers and fabric filters.
                                  108

-------
   450
   400
UJ
Qi




1
350
   300
   250
                          10         15         20        25

                                PRESSURE RISE - IN.WG
                                                                30         35
      Figure B-5.  Mixing section pressure rise  for ADTEC scrubber

                                  (1 kg H20/kg gas)
                                          109

-------
Table B-l.  DATA SUMMARY FOR CHROMASCO PILOT MODEL ADTEC SCRUBBER
System
Run Product Configuration
1 50% FeSi A
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
B
B
B
B
B
C
C
C
C
C
C
C
C
D
D
D
D
D
D
D
22 ^ D
23 507. FeSi D
24 757» FeSi D
25 ^ D

Water/Gas
0.97
1.4
1.3
1.3
0.91
0.72
0.80
1.21
1.21
1.40
1.21
1.36


0.94
0.82
0.74
0.70
0.64
0.88
0.79
0.80
1.24
1.26
1.22
Water
Temperature
(°F)
385
355
350
350
350
360
380
380
360
350
370
350


375
370
395
350
350
405
403
405
405
400
400
Outlet
Concentration
(g/m3)
0.046
Aborted
0.124
0.133
0.236
0.439
0.314
Aborted
0.277
0.169
0.227
Aborted
Inlet only
Inlet only
0.167
0.183
0.078
0.185
0.602
0.236
0.112
0.178
0.114
0.073
0.114
(grains /scf)
0.020
Aborted
0.054
0.058
0.103
0.192
0.137
Aborted
0.121
0.074
0.099
Aborted
Inlet only
Inlet only
0.073
0.080
0.034
0.081
0.263
0.103
0.049
0.078
0.050
0.032
0.050

-------
Table B-l.  (Concluded)


Run
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45

Product
75% FeSi
1
V
75% FeSi
50% FeSi
J
50% FeSi
CC70
CC70
SiMn






\






i
SiMn
System
Configuration
D
D
D
D
E
E
F
F
F
F
F
F
F
F
G
G
G
G
G
G

Water /Gas
1.36
0.95
1.21
1.15
1.13
1.16
1.07
1.07
1.05
1.20
1.13
1.08
1.08
1.17
1.08
1.24
1.72
1.2
2.25
1.78
Water
Temperature
(°F)
340
400
375
345
385
400
400
400
400
390
380
380
390
395
390
375
350
390
375
400
Outlet
Concentration
(g/m3)
0.046
0.066
0.069
0.078
0.378
0.286
0.192
0.201
0.22
0.23
0.03
0.08
0.048
0.11
0.032
0.055
0.08
0.016
0.018
0.005
( grains /scf
0.020
0.029
0.030
0.034
0.165
0.125
0.084
0.088
0.097
0.102
0.013
0.036
0.021
0.048
0.014
0.024
0.035
0.007
o'.oos
0.002

-------
     Table B-2.  PARTICLE SIZE DISTRIBUTION OF FERRO-SILICON DUST-i'



                                        80%                       50%
                                   Ferro-Silica              Ferro-Silica

Finer than 0.1 micron                 86.57                      89.97

Finer than 0.2 micron                 91.32                      92.47

Finer than 0.3 micron                 97.57                      96.63

Finer than 0.5 micron                100.00                     100.00
                                        112

-------
On the basis of currently available information, this system appears to
offer significant improvement in the collection of fine particles, at
modest energy consumption rates, where a waste gas is available which
contains a sufficient amount of thermal energy to heat the water.
                                   113

-------
Cross -Flow Nucleation Scrubber
Teller Environmental Systems
New York, New York

The cross-flow nucleation scrubber is a vertical packed bed of tellerette
packing with water flowing down through the packing.  Figure B-6 presents
a schematic diagram of the scrubber.  The quenched gas passes horizontally
through the packed section.  The nucleation system removes particulates
by a combination of particle growth via condensation, diffusiophoresis,
agglomeration, and inertial impact ion.  Condensation of water occurs
preferentially on submicron particles.  These particles not only increase
in size because of condensation, but collision between particles tends to
become less elastic, resulting in acceleration in growth via agglomera-
tion.  The diffusiophoretic contribution results from a molecular flow of
water molecules toward the cold particle surface.  This flow of water, in
turn, will sweep the submicron particles toward the cold surface.

The "angle of wetting" and the structure of the participate are reported
to be important factors in the design of each unit. I/  Pilot tests are
conducted on each source in order to determine the design factors that
must be used.  Quenching with water is reported to give better results
than does the addition of steam.  In some cases a small amount of sonic
energy upstream of the tellerette section is used to improve nucleation. 5/

The nucleation scrubber has been evaluated on a 15% slip stream from a
container glass furnace and on a full-scale basis on a fiberglass fur-
nace. 5j&J  Test results fox the container glass furnace are presented in
Table B-3, while Table B-4 presents the data for the fiberglass furnace.
Figure B-7 presents a schematic diagram for the installation at the fiber-
glass furnace.fi/  The source testing methods used to obtain the data re-
ported in Tables B-3 and B-4 are not known.  In addition, we were not able
to ascertain whether the testing was performed by an independent group or
by Teller Environmental Systems.  Although the particle size of the emitted
particulates was not reported in the above tests, other investigators have
established that the emissions from glass furnaces contain very high per-
centages of submicron particulate.2/

The range of operating conditions for the nucleation scrubber as reported
by Dr. Teller are as follows :V

Pressure drop - 2.5 cm - 10.2 cm H20 in packing
                0 cm - 20.3 cm H£0 in quench section
Water rate - 0.67-5.3 liters/m3 (5-40 gal/1,000 ft3) (not including
               quench)
                                    114

-------
                          -CONDENSATION ON
                           SUBMICRON PARTICLES
Ul
                                           PACKED
                                         "PARTICLE"
                                         GROWTH
                                         BECAUSE OF
                                         CONDENSATION
CLEAN
 GAS
                                                                "PARTICLE" REMOVED
                                                                BY INERTIAL  IMPACT
                                                                ON FILAMENTOUS PACKING
                            Figure B-6.  Schematic diagram of nucleation scrubber

-------
     Table B-3.  PERFORMANCE TESTS ON TELLER NUCLEATION SCRUBBER
                       (Container Glass Furnace)
                                        Inlet                    Outlet

SOX  (ppm)                               198                       4

Particulate (g/m3)                        0.42                    0.07
            (grains/scf)                  0.18                    0.03

Opacity (%)                              40                     - 5
System pressure drop:  28 cm of water
Efficiency on Particulates:  83.5%
      Table B-4.  PERFORMANCE TESTS ON TELLER NUCLEATION SCRUBBER
                         (Fiberglass Furnace)


SOx (ppm)
Gaseous fluoride (ppm)
Particulate (g/m3)
(grains/scf)
Boron particulate (g/m3)
(grains/scf)
Inlet
65
58
0.53
0.23
0.06
0.025
Outlet
< 0.3
0.3
0.02
0.01
< 0.001
< 0.0005
Opacity (7.)                              35                       5-10
Exhaust gas flow:  184 m3/min (6,500 scfm)
Liquid recycle:  1,323 liters/min (350 gpm)
System pressure drop:  10.2 cm (4 in.) of water
Efficiency on particulate:  95.6%
                                    116

-------
                                       •RECYCLE LIQUID
     QUENCH
             INUCLEATION
             I SCRUBBER
                                                                 COOLING
                                                                 TOWER
STACK
     NUCLEATION SCRUBBER
     SODIUM-CALCIUM LOOP
                                          CLEAN
                                          EXHAUST
BLOWER
                                          SOLIDS FOR
                                          POTENTIAL REUSE
                                          IN FURNACE
                                                                        PUMP
           Figure B-7.  Schematic diagram of installation of nucleation
                         scrubber systems on fiberglass furnace

-------
Superficial gas velocity - 2.1-2.4 m/sec (7-8 ft/sec)

Cost (FOB) - $70/m3/min up to » 85 m3/min ($2/cfm up to « 3,000 cfm)
             $1.05-$1.75/m3/min for higher cubic meter per minute
               ($0.30-$0.50/cfm for higher cubic feet per minute)

Available test data do not show particularly high collection efficiency
for small particulates, but the device has not been evaluated on a suf-
ficient number of sources to assess its overall performance capability.
Testing on a pilot-scale unit using EPA approved methods is recommended.
                                   118

-------
"Dynactor" Wet Scrubber
R. F. Industries, Inc.
Marlboro, Massachusetts
R. F. Industries has done limited development work on a wet scrubber sys-
tem that functions like a liquid driven eductor.  Figure B-8 is a cross-
section view of a single stage "Dynactor."2/  Liquid entering the system
at 40-200 psi is atomized into thin films and droplets of average thick-
ness or diameter less than 1/64 in. (0.04 cm or 400 microns).  This liquid
discharge diffuses or expands into the reaction chamber causing air or gas
to be aspirated by being trapped within the moving shower of films and
droplets.  The resulting mixed fluid then continues to travel down the
reaction column with intimate contact maintained between gas and liquid.

Little if any meaningful test data are available on this system and ap-
parently a small-scale demonstration unit is the only system currently
in operation at R. F. Industries.  Representatives of R. P. Industries
claim that the "Dynactor" has been tested at a plant producing catalyst
materials.  A "Dynactor" unit operating on the effluent from a cyclone
reduced the grain loading from an inlet value of 4.3 g/m3 (1.89 grain/
scf) to an outlet value of 0.02 g/m3 (0.01 grain/scf)--a collection effi-
ciency of 99.4%.  Although the particle size distribution of dust emitted
from the cyclone was not measured, it is reasonable to assume that a
major portion of the dust was less than 5 microns.  Water rates are re-
ported to be about 0.67 liters/m3 (5 gal/1,000 ft3).-/

The Control Systems Laboratory is currently funding additional testing
of this device.
                                   119

-------
Radial  Pressure
Transformation
Section
                                     Liquid Input,40 to 200  psi
       •Air Input, Low Velocity,
        Ambient  Pressure
High Velocity, Sub-ambient
Pressure

       Shower of  Thin Films and Particles
                                             Reaction Column


/in \\\\\
//I'M'
/ ' ' 1 1 \ *
/ / 1 ' ' \ V
'/ !U
/ / \\
i i i

t Gas Output
-1 	 Baffle
Reservoir/Separator
(Liquid)
1.
                                                                 Liquid Output

           Figure B-8.  Single-stage dynactor  system (cross-sectional view)
                                           120

-------
Pentapure™ Impinger
Purity Corporation
Elk Grove Village, Illinois

The Purity Corporation, Elk Grove Village, Illinois, has recently an-
nounced the development of an air pollution control device which is
reported to deliver high efficiency particle removal with low energy
requirements.£/

Figure 8-9 is a sketch of the Pentapure™ Impinger.  The agglomeration
duct has a diameter of 30.5 cm (12 in.) at the entrance and 15.2 cm
(6 in.) at the exit.  The length of the converging section (agglomera-
tion duct) is about 0.61 m (2 ft) and the distance from the exit of the
agglomeration duct to the impingement surface is around 23 cm (9 in.)

As the polluted air stream enters the Pentapure™ Impinger, a differential
velocity, which induces particle collision, is produced in the converging
section.  Submicron particulates will accelerate to a greater degree
within the converging section than the larger ones, causing a difference
in momentum which may result in collisions and agglomeration between the
different size particulates.

To further enhance this particle-on-particle target action, a medium of
larger particles can be injected upstream of the converging section.  The
most common substance is atomized water of controlled droplet size.

Collection of the particulate is achieved when agglomerate clusters im-
pact on the impingement surface.  Present models of the Pentapure™ Im-
pinger use either a fixed target or a moving belt onto which a water spray
can be directed to further enhance particle entrapment.   Reliable test
data are not available for this device.
                                    121

-------
                                                 Water In
N>
                                                                                     Clean Air Out
           Agglomeration Duct



                 Water In
              Atomizing Agglomeration Nozzle
                                                                       Coarse Impingement Spray




                                                                       Impingement Surface




                                                                                  Slurry Water Out
                        Figure B-9.  Schematic diagram of  Pentapure™ Impinger

-------
Mystaire™ Scrubber
Heat Systesm—Ultrasonics
Plainview, New York

A scrubbing system involving a wetted-knit mesh filter has recently been
developed by Heat Systems—Ultrasonics.  Figure B-10 presents a sketch
of a laboratory-scale system.  As shown in Figure B-10, a short spray
chamber proceeds the waterweb filter.  Gas and water flow is co-current.
A conventional mist eliminator is placed downstream of the waterweb to
collect any entrained water drops.  Test data are not available on this
device.
                                    123

-------
Inlet—~
         6"
        JJ
                                  Mist- Eliminator
                          Waterweb Filter
Drain  (1" NPT)
	30"	
                                            1/3"
                                    SIDE VIEW
                                                          Blower
                                                          Exhaust
                                It
                                                                Blower
                                                               II"-
                                                                                                      12" Dia.
                                                                                                        Water
                                                                                                        Inlet
                                                                                                        (1/2" NPT)
                                                        INLET END VIEW
                                                              TM
                Figure B-10.  Schematic diagram of the Mystaire   laboratory scrubber

-------
Dupont Wet Scrubber System
Dupont Company
Wilmington, Delaware

The Dupont Company has recently announced the development of a wet
scrubber system that appears to be nearly the same system as the Mystaire™
Scrubber.  Figure B-ll presents a schematic diagram of the Dupont de-
vice.  As  shown in  Figure B-ll the gas stream  is  first sprayed with water
and  then the gas-liquid mixture is passed through a  fiber bed structure
having a pore size  sufficiently fine and a path length sufficiently  long
to thoroughly contact  the gas stream containing the  particulate  solids
and mist with the liquid .IP./

Tests have been conducted on a laboratory-scale system employing a 10.1
cm (4 in.) diameter knitted fiber bed.  Overall bed depths of 10.1 and
20.2 cm were used.  Table B-5 summarizes the results of the test runs.!!/
Talc dust with the following reported size distribution was used in the
tests:  50% < 2.4 microns, 28% < I micron, and 14% < 0.5 micron.  Methods
used to determine the particle size of the talc dust were not reported
in Ref. 10.  The test results in Table B-5 show quite high collection
efficiencies with relatively low pressure drop.  Sampling procedures
used to obtain the data in Table B-5 were not reported.12/  Titanium
dioxide was also used as a test dust with substantially the same re-
sults .I2/

Tests were also conducted with different diameter fibers.  As shown in
Figure B-12, smaller diameter fibers were more efficient on an equal
weight per unit volume basis.  The packing density was 128 kg/m3 (8 lb/
ft3) and the void fraction was 0.983 for both beds.   The bed material
was 304 stainless steel in knitted form.  The specific surface area of
the 6-mil  fiber diameter bed was 459 m2/m3 (140 ft2/ft3) while that of
the 3.5-mil fiber diameter bed was 754 n^/m3 (230 ft2/ft3).  Improved
efficiency probably results from a longer path length in the fiber beds
composed of smaller diameter fibers.

Reference 12 indicates that the device can operate successfully at scrub-
bing liquid rates of 0.13-13 liters/m3 (1-100 gal/1,000 ft3) of gas con-
taining 22.9 g/m3 (10 grain/scf)  of dust.   At higher liquid rates,  dust
loadings in excess of 457.6 g/m3 (200 grain/scf) were handled.

In order to prevent plugging of the bed, the gas-liquid throughput must
exceed the flooding velocity.  Figure B-13 presents a generalized plot
of the flooding velocity in terms of important system parameters.  The
flooding velocity is a function of the characteristics of the fiber bed
structure,  the materials being handled, and the various flow rates.


                                    125

-------
                                             FIBROUS
                                             BED
                                             STRUCTURE
                 SLURRY
                 OUT   f
Figure B-ll.  Schematic diagram of Dupont wetted-fiber bed scrubbe
                                                                10/
                                  126

-------
Table B-S.  TEST DATA ON DUPONT SCRUBBER
Run
Ho.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Bed
(cm)
10.2
10.2
10.2
10.2
10.2
10.2
10.2
10.2
20.3
20.3
10.2
10.2
10.2
10.2
10.2
10.2
Liquid Spray Rate
(liters/
mln-m2 (gpm/ft2
Depth Bed Density of of
(In.) (kg/in3) (lb/ft3) Bed Material chamber) chamber)
4 64.1 4 Bed 896 22
4 composed
4 of
4 16 mil
4 diameter
4 - knitted 977 24
4 polyethylene
4
8
8
4 320.4 20 304 stainless steel 407 10
4 16-mll diameter
4 640.7 40 304 stainless 407 10
4 11-mll
4 192.2 12 diameter 407 10
4
Gas Throughout
(m3/mln)
5.04
5.89
1.30
1.84
2.27
1.36
2.66
5.04
1.36
2.61
3.71
5.10
4.11
5.01
3.88
4.96
(ft-Vmln)
178
208
46
65
80
48
94
178
48
92
131
ISO
145
177
137
175
Superficial
Gas Velocity
(m/sec)
10.4
12.2
2.7
3.7
4.9
2.7
5.5
10.1
2.7
5.5
2.7
3.7
3.0
3.7
2.7
3.7
(ft/see)
34
40
9
12
16
9
18
33
9
18
9
12
10
12
9
12
Inlet Dust
Concentration
(grains/
30
32
11
25
37
64
43
32
149
110
85
85
64
82
50
62
13
14
5
11
16
28
19
14
65
48
37
37
28
36
22
27
Pressure Drop
(cm H20)
28.7
34.8
7.4
8.6
11.4
10.9
20.6
33.5
22.9
48.3
9.4
12.2
24.4
25.1
27.2
26.2
11.3
13.7
2.9
3.4
4.5
4.3
8.1
13.2
9.0
19.0
3.7
4.8
9.6
9.9
10.7
10.3
Collection
Efficiency
98.9
98.7
97.7
98.4
99.2
99.8
98.8
98.9
99.4
99.6
98.8
98.5
98.6
98.9
99.4
98.7

-------
to
<
o
UJ
to
  o
  l/l
  to
  ••
 o

 i
      10,000
       1,000
         100
          10
         1.0
3.5 Mil Fiber

Diameter
                                              6 Mil Fiber

                                              Diameter
99.99
                                  99.9
                                  99.0
       U

       z
       LU

       y

       LJ_
       U.
       UJ



       O
                                         o
                                         u
                                  90.0
                                  0%
             024       6       8       10      12



                          THICKNESS OF  FIBROUS

                          BED STRUCTURE - INCHES


Figure B-12.  Measured collection efficiency of Dupont wetted-knit

                       mesh separator (2 micron particles)
                                    128

-------
               10°
    CM
 *
CN
N>
VO
CO
 W
  O
 O)
              10-1
              10-2
                                        G = Superficial Mass Velocity of Gas - lb/(Sec)(Ft2)
                                        9C = Gravitational Constant - (lb)(Ft)/Sec2 (Ib Force)
                                         L = Superficial Mass Liquid Velocity - lb/(Sec)(Ft2)
                                         S = Surface Area Packing - Sq Ft/Ft3
                                        V0= Superficial Gas Velocity - Ft/Sec
                                         * = Void Fraction of Structure -  Dimensionless
                                         L = Liquid Viscosity - Cent!poise
                                         P = Gas Density - Ib/Ft3
                                         \ = Liquid Density - Ib/Ft3
                             Flooding Velocity,
                             Fibrous Bed Structures
                        10"
   ID'2

_U_£_V/2
                                                                   10-1
                                                                                10°
                     Figure B-13.  Correlation of flooding velocity of Dupont  scrubber

-------
The Dupont scrubber system appears to be very similar to the Mystaire™
unit and additional testing with EPA approved methods is warranted.
                                    130

-------
                              REFERENCES

 1.   Sales Brochure, Lone Star Steel (1973).

 2.   Private Communication, Mr. T. W. Ewans, Lone Star Steel, (information
       obtained during visit to installation) (1973).

 3.   "A New Process for Cleaning and Pumping Industrial Gases—The Aronetics
       System," U.S. Patent No. 3,613,333.

 4.   Private Communication, Mr. H. E. Gardenier, Aronetics, Inc., Tullahoma,
       Tennessee, July 1973.

 5.   Private Communication, Dr. Aaron Teller, Teller Environmental Systems,
       November 1972.

 6.   Teller, A. J., "Control of Emissions from Glass Manufacture," Ceramic
       Bulletin.  £1(8), 637 (1972).

 7.   Stockham, J. D., "The Composition of Glass Furnace Emissions," Journal
       of the Air Pollution Control Association. 2J.(11), 713 (1971).

 8.   Private Communication, Mr. Stanley R. Rich, R. P. Industries, Inc.,
       Marlboro, Massachusetts, October 1973.

 9.   Pentapure™ Impinger. Purity Corporation, Elk Grove Village,  Illinois,
       Sales Brochure (1973).

10.   Lucas, R. L., and H. F. Porter, "Process and Apparatus for Wet Scrub
       Removal of Dust and Mist from Gases," U.S. Patent 3,370,401.
                                    131

-------
                               APPENDIX C

            CONDENSATION SCRUBBING FOR PARTICULATE COLLECTION

INTRODUCTION

The collection of submicron particulates from various gas streams can be
enhanced by the use of condensation effects.  Over a period of years,
there have been many references to the use of scrubbing systems in which
condensation effects have occurred.  In most of the early literature,
there appeared to be no deliberate attempt to optimize the effects of
condensation phenomena.  Available theoretical and experimental informa-
tion on condensation scrubbing is summarized in the following sections.
Specifically, the following sections  present:  (1) theoretical studies
of condensation phenomena; (2) experimental studies of particle condi-
tioning by steam and condensation scrubbers; (3) estimates of costs of
condensation scrubbers; and (4) areas and/or methods of potential applica-
tion of steam condensation.

THEORETICAL ASPECTS OF CONDENSATION SCRUBBING

Condensation scrubbing involves particle removal from the gas stream via
the use of a temperature gradient, a vapor concentration gradient, vapor
condensation, or combinations of the three.  It is not necessary that all
forces act simultaneously, but in most actual cases they do.

Forces which cause particle motion may result from momentum differences
imparted to the particle on opposite sides by the molecules colliding
with it or desorbing from it.  In the case of a temperature gradient,
hotter (and thus faster) molecules colliding with the particle will impart
a higher momentum to the particle than the cooler (slower) molecules.
Aerosol particles will then drift in the thermal gradient toward the cold
surface.  Particle movement by this mechanism is called thennophoresis.

In a concentration gradient, which is accompanied by diffusion but not
necessarily by net motion of the gas phase, the heavier molecules will
                                     132

-------
 impart a higher momentum than the lighter molecules.   If there is a net
 motion of the gas phase (Stefan flow),  additional force is  applied to
 the particles.  The combination of forces due to Stefan flow and the
 molecular weight or concentration gradient is referred to as the diffusio-
 phoretic force and particle motion as diffusiophoresis.

 Mass transfer, which causes the diffusiophoretic force, can be caused by
 evaporation,  condensation or a chemical reaction.   Particulate collection
 via thermophoresis and/or diffusiophoresis is most often considered in
 conjunction with the use of water as  a  scrubbing media; i.e.,  wet scrubbers.
 In the case of condensation of water  vapor,  the  Stefan flow force acts
 towards the condensing surface while  the molecular weight gradient force
 acts away from it.   The net, force is  toward  the  cold  surface.

 Particle growth resulting from the condensation  of vapor on particles
 may occur.    An advantage of particle growth would be improved particle
 collection  by inertial impaction.

 The individual mechanisms or forces which may be acting in  condensation
 scrubbing have been studied in considerable  detail from a theoretical
 standpoint.   Theoretical models  for thermophoresis and diffusiophoresis
 have been advanced  by  several  investigatorsil2/  (also see Appendix A).

 Nucleation  and condensation processes have been  studied in  some  detail.
 Water vapor may  self nucleate  and  condense into  drops  in the absence  of
 condensation  nuclei if the  degree  of supersaturation  is  great  enough.   At
 lower supersaturations,  however, a nucleating surface  is  necessary  for
 condensation.  The  conditions  necessary  for  condensation to occur on  a
 surface are determined by applying equilibrium thermodynamics.39/  A
 system is in  stable (unstable) equilibrium at  constant  temperature when
 the  Helmholtz  free energy is minimized  (maximized).  This fact is used  to
 find  the  equilibrium saturation ratio for  a  surface exposed to the vapor.
 (The  saturation ratio  is the ratio of the vapor pressure at the point in
question  to that of a  plane water  surface.)   If,  then, the saturation is
greater than required  for equilibrium, condensation occurs and will con-
tinue until equilibrium  is again attained.

If an  insoluble surface  is present, drops of the critical radius will
form on this surface at a higher rate than they will self nucleate.  This
is due to the fact that they have less volume and require fewer molecules
than for the case of homogeneous nucleation;  hence the free  energy increase
is smaller.   Their volume is smaller because they can form as a partial
sphere resting on the surface rather than the complete sphere necessary
in the absence of the surface.
                                    133

-------
If the particle is soluble, nucleation will occur even more readily.
Raoult's law states that the equilibrium vapor pressure over a solution
is less than that over pure water by a factor equal to the mole fraction
of water in the solution.

After a drop of critical size has been nucleated, it grows at a rate
determined by the ambient environment and the conditions at its surface,
just as the condensation flux to a flat wall depends on these boundary
conditions.  References 40-43 discuss nucleation, condensation, and
particle growth in detail.

Theoretical aspects of condensation scrubbing have also been investigated
to some extent.  Hales and his associates at Battelle Northwest have ex-
plored the possibilities of aerosol removal in an enclosed environment
at a condensing stream boundary.10-12/  These studies were concerned with
a nuclear reactor accident and their analyses were confined to liquid
sheets with only natural convection as the air moving mechanism.

Sparks and Pilatil/ published an analysis of a spray column with conden-
sation taking place on the drops.  Their analysis indicates that for the
case of Stokes Number equal to 0.6, the collection efficiency in a spray
column, where the vapor pressure gradient is 10^ mbar/cm, was 100% as com-
pared to an efficiency of 0% with no vapor gradient present.

The most comprehensive analysis of condensation scrubbing has been per-
formed by Calvert and his associates under contract to the Environmental
Protection Agency.iftliH.'* The objective of the research reported in
Ref. 16 was to determine the feasibility of using condensation scrubbing
for fine particulate collection.  Models have been developed for the pre-
diction of particle collection in scrubbers of several representative
types.

Models for plate scrubbers (as represented by sieve plates) and packed
scrubbers have been developed to the point where they appear to predict
with acceptable accuracy the simultaneous effects of heat and mass trans-
fer for gas/particles and gas/liquid diffusiophoresis, thennophoresis,
Brownian diffusion, inertial impaction, and changes in particle concentra-
tion, humidity, and temperature as the gas flows through the scrubber.—'

The model for spray scrubbers required appreciable effort on the computa-
tion of particle trajectories for the combined effects of flux forces
(i.e., thermophoresis and diffusiophoresis) and inertial collection by
*  Research is supported by the Control Systems Laboratory
                                     134

-------
liquid drops was found to be defined by a new dtmensionless parameter
which is the ratio of flux force deposition velocity to the gas velocity.
This ratio was called the "flux deposition number."  If the flux deposi-
tion number is greater than 0.1, particle collection efficiency will be
good and if it is less than 0.01, efficiency will be poor.  The model
for collection from an impinging jet showed that negligible benefit is
given by flux force deposition because of the very short contact time.—'

Theoretical analysis showed that, in general, the particle deposition
velocity has to be on the order of 0.1 cm/sec (0.04 in/sec) or larger
for appreciable collection efficiency to occur.  Diffusiophoresis can
produce deposition velocities this high under the heat and mass transfer
conditions of a realistic scrubber, but thermophoresis generally cannot.
If particle growth due to condensation occurs, inertial impaction of the
resulting 2 or 3 micron diameter particles can give sufficiently high
deposition velocities.  Particle growth has the advantage that the high
deposition velocity can persist after the heat and mass fluxes are dis-
sipated, because once the particles are enlarged, they can be collected
to an increasing degree at the cost of relatively little additional
pressure drop.M/

EXPERIMENTAL STUDIES OF CONDENSATION SCRUBBING

Experimental studies related to condensation scrubbing have been performed
on individual mechanisms (i.e., thermophoresis, diffusiophoresis, and
particle growth) and on entire scrubber systems.  Experimental studies of
thermophoresis are presented in Refs. 5 and 17-20.  The results of experi-
mental work on diffusiophoresis are presented in Refs. 21-23.

Experimental investigations on steam condensation related to particulate
removal from gas streams has been focused on two areas:  (1) particle con-
ditioning by steam condensation, and (2) performance of condensation
scrubbers.  Very little systematic work has been done in either area.

Table C-l presents a summary of the pertinent previous experimental work
conducted on particle conditioning by steam and on condensation scrubbing.
The particle conditioning experiments demonstrate that buildup of particles
(i.e., increase in particle size) by condensing steam can be achieved.
However, with the exception of the work of Prakash and Murray, particle
loadings used in the experimental work were very low and highly unrealistic
for most industrial situations.  Steam usage rates up to 1.0 kg/n»3 (62.5
lb/1,000 ft^) of gas were  used in the particle conditioning experiments
summarized  in Table C-l.  The amount of steam condensed was reported only
by Prakash  and  Murray.
                                    135

-------
                               Table C-l.  SUMKAHY OF SELECTED  EXPERIMENTAL INVESTIGATIONS  OF STEAM CONDENSATION
 Investigator

I. Particle
   Conditioning

   a.   Fahnoe
       et
                 Condensation Device
                 5.1 cm dia. standard
                 stainless steel pipe
b.  Goldsmith    5 cm dia., 50 cm
    and May.22'    long, water-cooled
                 glass tube
                                        Test Conditions

Aerosol:  HaCl,
silver iodide
Aerosol particle
size:
a) NaCl-0.80-1.2 nm
b) Silver iodide-
   maximum size
   0.8 urn
Aerosol grain
loading:
a) 0.001-0.002 g/

b) 0.0007 g/
   I0r
   (silver Iodide)
Steam Injection
rate:  0.16 kg/m3 of
air
Steam condensation
rate:  estimated at
0.01 kg/m3
of air
Inlet air temperature:
70-90'F
Relative humidity of
Inlet gas:   < 20%
Aerosol type,
aerosol particle
size
                                           Secondary Aerosol
                                           Collection System
a. Cyclone
b. Peabody
   scrubber
                                          Aerosol:  oxide
                                          particles
                                          Aerosol particle
                                          size:  0.05-0.5 urn
                                          Steam injection rate:
                                          0-1 kg/nj of gas
                    Gas velocity
                                          No secondary collec-
                                          tion system.
                                          However, elbow at
                                          exit of condenser
                                          provided Impaction
                                          surface
                                                                                                        System Performance
Condensation system:
99% buildup of
aerosol to particles
> 2 urn with HaCl, but
only 71% buildup with
silver iodide.
Secondary collection
system:
a) Cyclone collected
   94.3% of particles
   z 2 \im
b) Peabody scrubber
   collected 90.5%
   of particles
   2 2 urn
                                             Up to 98% removal of
                                             aerosol
                                                                                                                                      Comments
Very  low grain  loading
and soluble  particulates
tend  to minimize  quan-
tity  of steam required.
Grain loading not
realistic of industrial
conditions.
Difference in the ex-
tent  of particle
buildup is apparently
due to different  type
of particle  and smaller
size of the  silver
iodide aerosol.
                                             Removal efficiency varied
                                             with velocity of aerosol
                                             through the tube.  This
                                             effect Is probably due
                                             to the elbow at the  exit
                                             of condenser, i.e.,
                                             inertial impaction plays
                                             an important part in
                                             the sv'ste?..

-------
                                                            fablu C-l.  (Continued)
Invest lea tor Condensation Device
c. DavlD26'27^ Vertical 10. 2 -cm dia.
glass pipe, water-
cooled condenser
Test Conditions
Aerosol Nine
particle sub-
stances were
Secondary Aerosol
Parameters Studied Collection System System Performance
Aerosol type, aerosol No secondary col lee- Extent of particle
particle size, tlon system on con- removal varied with
aerosol grain loading denser particle size. Re-
Conments
Low grain loadings
tend to minimize
quantity of steam
                                        tested, Including
                                        examples of hydro-
                                        phobic, hydrophillc
                                        and hygroscopic
                                        particles.  These
                                        aerosols were tested
                                        over a particle con-
                                        centration range of
                                        102 to 2 x 106 par-
                                        ticles/cm3 and over
                                        a particle size
                                        range of 0.01 to
                                        2.7 urn In dia.
                                        Steam Injection rate:
                                        0-0.72 kg/m3 of air
d.  Prakash and
                 Variable  length glass
                 pipe of 2.5 cm I.D.
                 steam Injected through
                 a simple  nozzle Into
                 test section
                                            moval of particles
                                            as a function of
                                            aerosol type varied
                                            as follows:
                                            a. Hydrophobic par-
                                               ticles-  up to
                                               90% depending
                                               on initial particle
                                               concentration
                                            b.  Hydrophlllic par-
                                                ticles:  up to
                                                987. depending on
                                                initial particle
                                                concentration
                                            c.  Hygroscopic par-
                                                ticles:  In excess
                                                of 991
                                                                                                                               required.
                                                  CaCO-
  3,
and
Aerosol:
ZnO,
Na2C03
Aerosol particle
size:  See
Figure C-l
Aerosol grain
loading:  0.7-7.0 g/
m3
Steam Injection
rate:  0.03-0.1 kg/
nr
Steam condensation
rate:  0.01-0.03 kg/
in
Gas  flow rate'  0.13 m3/
mln
Aerosol type,  aerosol
particle size,
aerosol grain loading
No secondary collec-
tion system
Particle growth was
found to depend upon
particle size, par-
ticle type, and par-
ticle grain loading
in some cases.  Par-
ticles soluble In
water or easy to wet
showed higher growth
rates.  For these
aerosols, particle
growth was not affected
by particle loading
or by the size of the
particles ulthln the
range studied.  Growth
rate  for particles in-
soluble In water Is
strongly dependent
upon  particle loading
and particle size.
Condensation was
favored on larger par-
ticles and bv higher
loading*.
Results of program
Indicate that par-
ticle conditioning
by steam condensation
can be of considerable
value In improving
efficiencies of gas
cleaning equipment.
Steam usage rates are
among lowest reported
In literature.

-------
                                                                     Table C-l.  (Continued)
        Investigator

       II. Condensation
           Scrubbing

          a.   Lohs"/
                          Condensation Device
                          Steam tower
00
          b.
              Schauer3^
Convergent-divergent
nozzles
                                                 Test Conditions
                                                                     Parameters Studied
                                                 Aerosol:
                                                 Suspension A:
                                                 Sodium sulfate
                                                 (soluble In HjO),
                                                 particle size
                                                 0.3-2.0 un
                                                 Suspension B:
                                                 Polystyrene
                                                 (hydrophoblc dust),
                                                 particle size 0.3-
                                                 2.0 urn
                                                 Steam Injection rate:
                                                 0.5-0.6 kg/hr
                                           Aerosol type
                                                                                            Secondary Aerosol
                                                                                            Collection System
                                          Spray tower with
                                          water rates ranging
                                          from 0.25-1.0
                                          llter/mln
Aerosol:  OOP
Aerosol grain
loading: 0.27 g/m3
Gas flow rate:
0.28-0.42 m3/nln
Steam Injection rate
3.2 kg/m3
Nozzle configuration
Double convergent-
divergent nozzle
system Is necessary
to  get good conden-
:  atlon
Pease-anthony
cyclonic scrubber
                                                                                                                  System Performance
                                             Suspension A:  971
                                             removal at 0.5 )>m
                                             with 3.3 Ib steam/hr
                                             Suspension B:  801
                                             removal at 0.5 urn
                                             with 1.5 kg steam/hr
Efficiencies on the
order of 99.91 were
obtained on 0.3 urn
particles
                                                                                                                                               Comments
                                             Separation efficiency
                                             of spray tower was
                                             enhanced by addition
                                             of steam.  Improvement
                                             of up to 701 was ob-
                                             served.  Major Improve-
                                             ment occurred In 0*3-0.5
                                             urn particle size range.
Very low grain loading
used In these experiments
tends to minimize quan-
tity of steam required.
However, the use of  a
hydrophobia dust
probably compensates.
              Lancaster3-^ Rectangular duct
                                                  Aerosol:  zinc
                                                  oxide
                                                  Aerosol particle
                                                  size:  ~  1 |im
                                                  Aerosol grain
                                                  loading:  0.23-
                                                  11.4 g/m3
                                                  Steam  Injection rate:
                                                  0.02-0.18 kg/in3
                                                                     Residence time
                                                                  Cyclone
                                                                Up to 95% removed
                                                                at grain loadings of
                                                                11.4 g/m3.  Up to
                                                                85% removal at grain
                                                                loading of 0.23 g/m3
                                                                    Inefficient  use was made
                                                                    of available steam
                                                                    because the  condensable
                                                                    fraction of  the Injected
                                                                    steam was condensed on a
                                                                    small percentage of the
                                                                    total particles In the
                                                                    system.

-------
                                                             Table C-l.  (Concluded)
Investigator
  d.
      Rosen and
      Kostlnli/
Condensation Device

Perforated plate
column with alternate
hot and cold plates
  e.   Lltvlnov33"36/
                  Venturl Scrubber
  f.
      lexebenln
      and
Netted wall
rectangular ducts
Teat Conditions

Aerosol:  Oil mist
Aerosol particle
size:  ~ 0.3 urn
Parameters Studied

Plate temperature,
and nuaber of pairs
of hot and cold
plates
Secondary Aerosol
Collection System

Hone
Aerosol:   Carbon
black and Apatite
Aerosol particle
sIze:   Carbon
black ~ 1 urn,
Apatite ~ 1.7 urn

Aerosol:   Tin fume
Aerosol particle
size:   ~ 0.1 urn
                                                                                    Two stage sieve
                                                                                    column
                                                                                    None
System Performance

Up to 90% removal on
fourth pair of plates
for a temperature
difference of 50°C
between hot and cold
plates
                                                                                       Collection efficiency
                                                                                       Increased under con-
                                                                                       ditions favoring con-
                                                                                       densation
      Comments

Experimental data
could be correlated
by equation of form
P = 12.5 q'0-56
where P = particle
penetration and
q ° mass of steam
condensed per gram
of Inlet particles
                                                                   Particle collection
                                                                   attributed to growth
                                                                   by condensation and
                                                                   sedimentation of par-
                                                                   ticles  by dlffuslo-
                                                                   phoretlc force.

-------
l.Or-
                              0.5
1.0                          5.0
    Particle Diameter, Microns
10
43  50
                     Figure C-l.  Cumulative mass fraction distribution of experimental aerosols used
                                   by Prakash and Murray (see Table C-l for summary of work)

-------
Frakash and Murray reported that within the range of their experiments,
the amount of steam condensed was found to be proportional to the amount
of steam injected--approximately 33% of the steam condensed.  The por-
tion of the steam condensed on the particles was not reported, but it
was assumed to be proportional to the total quantity of steam condensed
in the system.

The main conclusions from the particle conditioning experiments are:

1.  Growth of particles is dependent upon type of aerosol.

2.  Dependence of particle growth on particle size and grain loading
varies with type of particle.

3.  Particles soluble in water or easy to wet were not strongly affected
by particle loading or particle size and exhibited higher growth rates.

4.  Growth rate of particles insoluble in water is strongly dependent upon
particle loading and particle size.  Condensation increased in gas streams
containing larger particles and higher grain loadings.

5.  Particle conditioning by steam condensation can be of considerable
value in improving the efficiencies of gas cleaning equipment.

Experimental work on condensation effects in scrubbers is currently being
conducted at Stanford Research Institute and at A.P.T., Inc., under con-
tracts from EPA.* Previous work by SRI investigators has shown that less
power input is required for the same efficiency when gases are cooled below
the dew point.  The current work at SRI involves an investigation of the
reduction in power input that can be realized in venturi or orifice type
scrubbers by taking advantage of condensation effects.  The purpose of the
work is to compare efficiency results at fixed power inputs (AP), with and
without condensation.

Experimental work at A.P.T., Inc., is directed to the study of plate-type
scrubbers.  Theoretical work by A.P.T., Inc., had indicated that plate-
type scrubbers should be good for condensation scrubbers on the basis of
their assumed heat and mass transfer characteristics.  Initial studies
have been performed with a single sieve-plate column.1^'   The test sec-
tion consisted of a 10.2 cm diameter glass column and the sieve plate was
made out of 1.6 mm aluminum sheet and had an overall diameter of 10.2 cm.
The plate had 30 perforations of 4.8 mm diameter, adding up to 6.67. free
   Work at both locations is being funded by the Control Systems
     Laboratory.

                                   141

-------
area on the plate.  Two straight weirs, 8.1 cm long and 9.5 mm high were
provided for liquid inlet and outlet on the plate.  The results of a
limited number of tests indicate that diffusiophoresis is the major con-
tributor to particle collection for a single plate.

Future experimental work at A.P.T., Inc., will focus on multiple plate
columns.  Multiple plate columns should provide staging effects which
permit better utilization of condensation effects.M/

The possibility of using multiple plate columns in conjunction with steam
injection was discussed by Bralove.     '  Bralove discusses the adaptation
of a Peabody perforated plate scrubber for steam injection (see Figure
p.-9) .44-467  with the single steam injection stage before the scrubber as
shown in Figure C-2, the action of the five plates would be to remove the
particles which were of the greatest size on the first plate, with the
efficiency of each plate rapidly diminishing.  The overall efficiency of
removal would not be significantly higher than the efficiency with which
the particles were preferentially wet.

Bralove suggests that the overall efficiency could be improved by increasing
the number of steam-injected stages by introducing points of injection
between each plate.  This would serve to wet the particles which were not
wet by the previous stage, this injection occurring after the wetted par-
ticles were removed.  Interstage cooling would be necessary for condensation
to take place in the next stage.

Figure C-3 shows Bralove's projected application of a typical plate scrubber
employing five stages of steam injection.  Steam is injected through ring
injectors into the gas stream before it enters each of the plates.  The
dust-laden gases pass through the orifices in the plates located directly
above.  The gases are then passed through an interstage cooler.  While
primary cooling occurs on the plate itself, additional cooling is necessary
to effect efficient condensation in the succeeding stage.  Instead of an
interstage cooler, an additional scrubber plate designed specially for
cooling might be used.  Based on limited experimental data for radioactive
dusts, Bralove estimated 85-90% cleanup per stage employing steam injection
and an overall efficiency exceeding 99.99%.

Bralove also discusses sudden cooling to effect condensation of water about
dust contained in a hot saturated gas Ji§/  A plate scrubber was suggested
by Bralove as being suitable for particle collection by sudden cooling Jt§'
                                    142

-------
                    Impingement
                    Baffle Plate
                      Top View
                   Cross Section
                Dirty Gas Inlet
                         Steam Injector
                         Ring
                                                  Cold Clean
                                                  Gas Outlet
                                                              \
                                                               I
                                                              r
                                                                             Cold Water
                                                                             Inlet
Water
Discharge
Figure C-2.  Five-stage Peabody perforated plate  scrubber adapted for
                          steam injection
                                     143

-------
Cold	*•
Collection Water
Dust-Laden
Gas  Inlet  ~~"
   Steam
   Injection
   Distributors/
Atmospheric
Steam.
                                                                         Interstage
                                                                         Coolers
                                                                                     To  Freon
                                                                                     Cooling
                                                                                   :  System
                                                                     Hot Dust-Laden
                                                                     Water to Disposal
                                                                     or Recircu lotion
                                                                     System
Figure  C-3.   Five-stage steam injection plate scrubber
                                   144

-------
Some of the major observations from previous experimental studies are:

1.  Collection efficiency depends heavily on the amount of water vapor
condensed per unit mass of dry gas.

2.  Collection efficiency also depends significantly on particle concen-
tration.  There is a trend of efficiency increasing as particle number
concentration decreases.

3.  Effectiveness of condensation scrubbing depends upon the type of dust
being scrubbed.  Efficiency varies in the following order:  hygroscopic
dust > hydrophilic  dust > hydrophobic dust.

4.  Mechanism of particle growth via condensation is dependent upon type
of particle.  The difference  in the growth mechanisms probably accounts
for some of the variations noted in scrubbing efficiency.

5.  The effects of scrubber design include the following:

a.  Multiple-stage or  continuous contact type equipment is superior to
single-stage condensation because it provides more opportunity for the
collection of particles after they have grown.

b.  Distribution of the condensation over several stages is preferable
because of the enhanced growth which can occur after the particle con-
centration has been reduced.

ECONOMICS OF CONDENSATION SCRUBBING

The economics of condensation scrubbing are expected to be dependent upon
the source and the characteristics of the emitted particulates.  Costs
for condensation scrubbing will be highly dependent on the amount of steam
which  is  condensed, especially if  steam or  fuel  to evaporate water must be
purchased.  Cooling water may also be required to condense the steam and
the amount of cooling  water needed will depend on the temperature rise of
the water in the  scrubbing system.  The efficiency of collection as noted
in the preceding  section varies with particle type and grain loading.  As
a result,  the cost  to  achieve a  specific degree  of' control will  vary with
those  parameters.

A general economic  analysis of condensation scrubbing has been performed
by Calvert,  et al.16»38/  The main conclusions  from the  analysis were:
                                    145

-------
1.  The most favorable situation occurs when the enthalpy of vaporization
is available from the gas to be cleaned.

2.  The purchase of steam might be justified when high efficiency collec-
tion of fine particles is required.

3.  Cooling water costs can be significant depending upon the system.

CONCLUSIONS

The work on condensation scrubbing which has been conducted to date in-
dicates it may be an important avenue to exploit in order to collect fine
particulates.  Two avenues appear most feasible:  (a) injection of steam
into a cold dust-laden gas; and  (b) sudden cooling of a hot and humid
gas.  Sudden cooling is an inefficient method for producing condensation
since most of the water condenses on the cooling surface, but enough may
condense on particles to affect particle collection.  A gas above 120°F
cannot be treated very effectively by steam injection because condensation
will not be significant.

Available theoretical and experimental data suggest that multiple stage or
continuous type contact equipment is superior to single-stage equipment
for both injection and cooling.  At present, there is little more than an
order of magnitude quantitative understanding of the phenomena occurring
when steam, water droplets, and particulates are mixed—especially in
equipment configurations that are likely to be useful in actual industrial
gas cleaning applications.

In order to resolve the areas of uncertainty and to develop reliable data
for engineering design, additional development effort is necessary.
Effort should be focused on pilot scale work in order to develop data
under realistic heat and mass transfer conditions.  The importance of par-
ticle properties and concentrations also indicates that the effects of
these parameters should be defined in more detail.
                                     146

-------
                                REFERENCES

 1.   Waldmann,  L., A.  Naturforach.  14A,  589 (1959).

 2.   Epstein, P.  S.,  Z.  Phys..  54,  537  (1929).

 3.   Brock,  J.  R., J.  Colloid Sci.. 1£,  768 (1962).

 4.   Jacobsen,  S., and R.  Brock, Jr., J. Colloid and Interface Sci.. 20,
       544 (1965).

 5.   Byers,  R.  L., and S.  Calvert,  Ind.  Ene. Chem. Fund.. 8, 646 (1969).

 6.   Facy, L.,  Arch.  Met.  Geophys.  Bioklim. 8,  229 (1955).

 7.   Schmitt, K.  H.,  and L. Waldmann,  Z. Naturf.. ISA, 843 (1960).

 8.   Deryagin,   B. V., and S. S. Duklin, Dokl.  Akad. Nauk.. SSSR (Phys.
       Chem.)  106, 851 and 111, 613 (1956).

 9.   Deryagin,  B. V.,  and  S. P. Bakanov, Dokl.  Akad. Nauk.. SSSR (Phys.
       Chem.)  147.  139 (1962).

10.   Hales,  J.  M., T.  W. Horst, and L.  C. Schwendiman, "Aerosol Transport
       in a  Condensing-Steam Boundary Layer," BNWL-1125, Battelle Northwest
       Report,  June  1, 1970.

11.   Hales,  J.  M.,  "Aerosol Transport in a Condensing-Steam Boundary Layer,"
       BNWL-SA-2359,  February 20, 1969.

12.   Horst,  T.  W.,  "A Review of Particle Transport in a Condensing Steam
       Environment," Battelle Northwest Report BNWL-848, June 1968.

13.   Sparks, L. E.,  and M. J. Pilat, Atmos Environment. 4, 651  (1970).

14.   Calvert, S. et al., Scrubber Handbook  (EPA Contract No. CPA-70-95)
       APT,  Inc., Riverside, California, August 1972.

                                     147

-------
 15.   Goldshmid,  D.  Leith,  Jr.,  and S.  Calvert,  "Flux Force Scrubbers -
       An Engineering Analysis," presented 164th ACS National Meeting,
       New York,  New York,  August 27,  1972.

 16.   Calvert,  S.  et al., "Feasibility  of Flux Force/Condensation Scrubbing
       for Fine  Particulate Collection," EPA Report  EPA-650/2-73-036 (EPA
       Technology Series).

 17.   Saxton, R.  L.,  and W.  E. Ranz,  J. Appl.  Phvs..  23,  917 (1952).

 18.   Schadt, C.  F.,  and R.  D. Cadle, J.  Phya. Chem..  65,  1689 (1961).

 19.   Rosenblatt,  P.,  and V.  K. LaMer, Physical  Review. 70, 385  (1946).

 20.   Fulford, G.  D., M. Moo-Young, and M. Babu, Canadian Journal of
       Chemical Engineering. 49,  553 (1971).

 21.   Davies, C. N., Aerosol  Sciences. Academic  Press, New York.

 22.   Meisen, Borkowicz, Cooke, and Garkas, Canadian Journal of Chemi-
       cal Engineering. 19,  449-457  (1971).

 23.   Postma, A. K., and J. D. McCormack, Nucl.  Teehn.. 10, 511 (1971).

 24.   Fahnoe, F.,  et al., "Aerosol Build-Op Techniques," Industrial and
       Engineering Chemistry. 43_ (6) pp. 1336-1346 (1951).

 25.  Goldsmith, P., and F. G. May, Aerosol Science. Ed. by C.  N. Davies,
       Academic Press, New York  (1966).

 26.  Davis, R. J., and J.  Truitt, "The Function of Condensing Steam in
       Aerosol Scrubbers," ORNL-4654, Oak Ridge National Laboratory,
       March 1971.

 27.  Davis, R. J., and J.  Truitt, "Additional Studies of the Function of
       Condensing Steam in Aerosol Scrubbers," draft copy.

 28.  Prakash, C.  B., and F. E. Murray,  "Particle Conditioning  by Steam
       Condensation," Department of Chemical Engineering, University of
       British Columbia, May 1973.

29.  Lohs, W.,  "Manufacture of Aerosols and Separation of Ultrafine Dusts
       in Spray Washers,"  Staub-Reinhalt Luft. 2£ (2) pp. 43-48,
       February 1969.
                                    148

-------
30.  Schauer, P. J., Industrial and Engineering Chemistry. 43, 1532 (1951).

31.  Lancaster, B. W., and W. Strauss, "A Study of Steam Injection into
       Wet Scrubbers," Ind. Eng. Chem. Fundamentals. 1£ (3) 362-369 (1971).

32.  Rozen, A. M., and V. M. Kostin, "Collection of Finely Dispersed
       Aerosols in Plate Columns by Condensation Enlargement," International
       Chemical Engineering. £, 464-467, July 1967.

33.  Litvinov, A. T., Khimicheskayer Promyshlennost. j3, 64-69 (1964).

34.  Litvinov, A. T., Stal. 7, 667-669 (1965).

35.  Litvinov, A. T., "Influence of Condensation on the Effectiveness of
       Capture of Fine Particles During Cleaning of Gases by the Wet
       Method," Zhurnal Prikladnoi Khimii. 40 (2) 353-361 (1967).

36.  Litvinov, A. T., "Fine Gas Scrubbing to Remove Highly Dispersed
       Hydrophobic Particles Using the Condensation Effect," Stal. 2,
       184-186  (1972).

37.  Terebenin, A. N., and A. P. Bykov, Zhurnal Prikladnoi Khimii  (USSR)
       45, 1012 (1972).

38.  Private Communication, Dr. Seymour Calvert, A.P.T., Inc., Riverside,
       California, December 1973.

39.  Abraham, F. F.,  "A Re-Examination of Homogeneous Nucleation Theory:
       Thermadynamic  Aspects," J. Atmospheric Sci.. 25j 47-53 (1968).

40.  McDonald,  J. E., "Homogeneous Nucleation of Vapor Condensation.
       II.  Kinetic Aspects," Am. J. Phys.. 31, 31-41  (1963).

41.  Fletcher,  N. H., "Size Effect in Heterogeneous Nucleation,11 J. Chem.
       Phvs.. 29, 572-576  (1958)  (see, also, 31, 1136).

42.  Fletcher,  N. H., The Physics of Rain Clouds. Cambridge University Press,
       London,  pp. 58-61 (1962).

43.  Howell, W. E., "The Growth of Cloud Drops  in Uniformly Cooled Air,"
       J. Meteorology. 6,  134  (1949).

44.  Bralove, A. L.,  "Radioactive Dust Separation Equipment - I,"
       Nucleonics. £(4),  37-40  (1951).
                                    149

-------
45.  Bralove, A. L., "Radioactive Dust Separation Equipment - II,"
       Nucleonics. £(5) 60-67 (1951).

46.  Bralove, A. L., "Radioactive Dust Separation Equipment - III,"
       Nucleonics. £ (6), 15-23, 33 (1951).
                                    150

-------
                            APPENDIX D

    CHARGED DROPLET SCRUBBERS AS PARTICIPATE CONTROL DEVICES

 INTRODUCTION

 Control devices using charged drops as the collecting bodies are cur-
 rently under study by several investigators-.-^—^'* In these devices, the
 particulates contained in a dirty gas stream are charged and collected
 on  the surface of droplets having an opposite  charge.  The size of the
 collection drops is chosen so that they can be readily removed by con-
 ventional control devices or will precipitate  in their space-charged
 field.  Theoretical aspects and limited experimental data on charged
 droplet scrubbers are reviewed next.

 THEORETICAL ASPECTS

 No  complete theoretical analysis of charged droplet systems is available
 at  this time.  The earliest work on charged droplet systems appears to
 be  that reported by Penney.—'  Kraemer and Johnstone discuss an electri-
 fied wet scrubber as part of their work on the collection of aerosol
 particles in the presence of electrostatic fields.£/  They present a
 general theoretical solution and experimental verification for the depo-
 sition of aerosol particles from a moving gas stream and suggest that an
 electrified wet scrubber would be one approach to the use of electro-
 static forces for particle collection.

 The use of charged drops to collect or agglomerate fine particulates has
 been reviewed in detail by Melcher and Sachar.—'  They consider mechanisms
 using injected droplets which fall between those of inertial scrubbers
 and electrostatic precipitators.  Their classification is shown in Figure
 D-l.  Categories I and II are mechanical and electrical scrubbers.  It
 is  interesting to note that Melcher and Sachar find the electrical
 scrubber to be roughly competitive with the mechanical scrubber in water
usage,   particularly for fine particles.   Energy requirements were not
*  Work in Refs. 2 and 4 is being supported by EPA,  Control Systems
     Laboratory.

                                   151

-------
                          Uncharged
                         Charged; No
                         Ambient  E
Charged,
Ambient E
Uncharged
I. Mechanical
scrubbers; inertial
impact from drop
injection.
                           W
IV. Collection
on half-surface
of polarized drop
having no net
charge.
Charged; no
ambient E
                         III. Drops having
                         net charge collect
                         particles having net
                         opposite charge.
Charged, and
ambient E
 II. Electrical scrubbing;
inertial  impact through
electrical propulsion of
drops.
                                                                          V.  Combination
                                                                          of  II, III and IV
                                                                                o
      Figure D-l.   Classification  of particle-drop interaction mechanisms!/
                                   (Melcher  and Sachar)
                                               152

-------
 evaluated.   Classes  III  and  IV consist of two  basic  agglomeration mecha-
 nisms  for charged particulates.   Class V is  a  hybrid category which com-
 bines  mechanisms  II,  III,  and IV with a conventional electrostatic
 precipitator.

 Melcher and Sachar also  discuss  simple models  for charged drop-particle
 interactions corresponding to the various classifications shown in
 Figure D-l.  The  purpose of  the  development  of models was to highlight
 the dependence of the various collection mechanisms  on  system parameters.
 Attention was  focused on Class III interactions  since they appear to
 hold the most  promise for a  new  approach to  particulate control.U

In subsequent work, Melcher,  et  al., identified three critical time
constants which they feel characterize the limitations of using charged
drops for collecting small charged particles.JLii2/  First, there is the
time  TR = e0/NQ$D ,  where in MRS units  co = 8.85 x 10'12 , N  is the
droplet density,  Q  is the charge per drop and  $D  is the mobility of
the drop.  If there is only one polarity of charged drops, this time
constant represents a time for these drops to be electrostatically
self precipitated on the walls of the channel or containers.  If the
system is composed of both positive and negative drops,  then the same
time constant (but based on one of the two families of particles) repre-
sent self agglomeration or at least electrical discharge among the drop-
lets.!/

A time constant playing a similar role for the fine charged particles
(which are to be removed by collection on the drops)  is defined as
T* = e0/nq$f  where  n,  q, and $f  are respectively defined as the num-
ber density, charge per particle, and mobility of the fine particles.
Again, in the case of unipolar charging, this is the  typical time for
self precipitation of like-charged fine particles and with bipolar charg-
ing it is typical of the time required for self agglomeration.  Finally,
there is a third characteristic time  Td = eo/NQ$f  which typifies the
time required for the fine particulate to be collected by or at least be
discharged by the droplets.

Melcher states that one of the most fundamental limitations on what can
be done using charged drops to collect fine charged particles comes from
the fact that  TR  is typically much shorter than  Td«  This is true be-
cause the mobility is generally proportional to the size of the particle
and hence (L » 0f. Thus, the charge from the drops will be available
to cause the collection of the small particles only for a time which is
short compared to the time actually required to collect a significant
fraction of the small particlesJj
                                    153

-------
Professor Pilat at the University of Washington has also performed some
theoretical analyses of charged droplet scrubbers as a part of his work
on these systems.  The details of his analysis were not available to
use in this report .V

Professors Donald Hanson and Charles Wilke at the University of California,
Berkeley, have investigated space-charge precipitators augmented by water
droplets.  Basically, the space-charge precipitator developed at the Uni-
versity of California consists of two parts:   (1) a short section where
both the particles and the added drops are charged by high voltage
coronas; and  (2) a section of grounded tubes or plates on which the parti-
cles and drops are collected.  In the collecting section, the drops and
particles migrate  to the surfaces, where they coalesce and flow from the
precipitator  as a  slurry.

Both theoretical and experimental studies on space-charge precipitation
have been conducted at the University of California.  Theoretical studies
have focused  on an analysis of the nature of drop and particle charging
and the analysis of drop production.  An engineering design for a pilot-
scale space-charge precipitator was also proposed.

EXPERIMENTAL  STUDIES OF CHARGED DROPLET SYSTEMS

Only limited  experimental work on charged droplet systems has been con-
ducted to date.  Professor Pilat at the University of Washington has a
pilot plant version of an "Electrostatic Droplet Spray Scrubber" in
operation.I/  In this device the droplets are charged positively and the
particulates  negatively.  The efficiency of the scrubber was increased
(when the charges were added) from 68.8% to 93.6% for 1.05-micron DOP
droplets (standard deviation of 2.5) and from 35% to 87% for 0.3-micron
droplets.  The scrubber requires about 0.7 liters of water at 100 psi/m-V
min (5 gal at 100 psi/1,000 cfm) and a small amount of power.!/  It is
interesting to note that the efficiencies reported by Pilat are much
higher than would be predicted by the theoretical models proposed by
Melcher.  The pessimistic position expressed by Melcher regarding charged
droplet scrubbers would appear to require further investigation.

TRW, Inc.,  is reported to have a proprietary charged droplet scrubber
which they are now developing.   This device is claimed to have an ef-
ficiency comparable to a venturi scrubber, but to require only about
0.02 hp/m^/min (0.5 hp/1,000 cfm).  A unit is currently in operation in
Japan.  Reliable efficiency data for fine particulates are not available.—'
The Control Systems Laboratory is currently funding additional work on
charged droplet scrubbers at TRW, Inc.
                                     154

-------
 Experiments on a laboratory-scale space charge  precipitator were  con-
 ducted at the University of California.  The  experimental  unit, depicted
 in Figure D-2, handled 0.85 m3/min of  air on  a  dry basis at a  velocity  of
 3 m/sec.   The fog injection system and the corona charging system were
 built according to specifications developed in  the engineering design cal-
 culations ..LJLi'   The  particulate material injected  included,  during
 various experiments, tobacco smoke,  oil mist, and several  other dusts.!'

 Tables D-l and D-2 show the extent of  precipitation  of  iron oxide dust
 with and  without the injection of fog.  Table D-l shows that about  92%
 of the iron oxide was  precipitated by  normal  electrostatic precipitation
 when no fog was added.   Table D-2 shows 100%  precipitation of  the iron
 oxide, within the limits of detectability,  when fog  was added.  The
 colorimetrie data shown in  Table  D-2 were reported to be accurate to
 within about 1% on the percent precipitation, and results  in Table  D-2
 should, therefore, be  interpreted as 99+% precipitation.!^'   The simple
 experiments summarized in Tables  D-l and D-2  probably do not accurately
 model a practical system.   However,  they do suggest  that charged  droplets
 have some potential for particulate  collection.

 Energy requirements for charged droplet systems are relatively uncertain.
 Kostow'al'  calculations yield a total energy requirement of less than
 0.02 hp/m3/min  (0.5 hp/1,000 cfm).  Melcher and Sachar2./ do not calculate
 energy requirements.  If we assume that a charging and collection system
 for  their Types III or IV mechanisms has energy requirements similar to
 that for a  conventional electrostatic precipitator,  we can estimate the
 electrical  power at less than 0.04 hp/m3/min (1.0 hp/1,000 cfm).   Melcher
 and Sachar's model suggests  water mist loadings of less than 0.02 kg/
m3/min (1 lb/1,000 cfm), which should require less than 0.04 hp/m3/min
 (1.0 hp/1,000 cfm) to produce.  Thus, total energy requirements of about
 0.08 hp/m3/min  (2 hp/1,000 cfm) appear likely for charged  droplet systems.

 CONCLUSIONS

 The currently available data on electrostatic scrubbers and related de-
vices  using charged drops indicate some potential for the collection of
 fine particles.  However, adequate data are not now available for a com-
prehensive  analysis of any of the existing devices or of the basic
mechanisms  involved.  While the ultimate efficiency and usefulness of
 these  devices cannot be determined accurately at this time, it should be
noted  that  the major workers in this field estimate,  from their experience
and analysis, that efficiencies comparable to, but not much better than,
conventional electrostatic precipitators can be obtained for submicron
particles.3»7>10/
                                     155

-------
                                                          TO VENT
TEMPERATURE
INDICATOR
                                GLASS
                                SECTIONS
                SAMPLING
                FILTER—*-
               PARTICLE
               INJECTION
               SYSTEM
                  CORONA
                  DISCHARGE
                     FOG
                     INJECTION
                     SYSTEM
             D.C. POWER
             SUPPLY
                            STAINLESS STEEL
                            TUBE
                                        ORIFICE
                                        METER
         DRAIN
STEAM
INLET
Figure D-2.  Experimental apparatus for space-charge precipitation
                                 156

-------
      Table D-l.  IRON OXIDE DATA, NO FOG ADDED TO THE  PRECIPITATOR
                    Corona
                      Off                  On                 Precipitation
Run                   "A"

  1                  0.900                0.024                    97.3
  2                  0.670                0.104                    84.5
  3                  0.546                0.019                    96.5
  4                  0.695                0.056                    91.9
  5                  0.715                0.068                    90.5
  6                  0.945                0.196                    79.8
  7                  0.620                0.008                    98.7
  8                  0.229                0.010                    95.6
  9                  0.222                0.012                    94.5
 10                  0.168                0.004                    97.6
 11                  0.215                0.004                    98.1
 12                  1.00                 0.202                    79.8
 13                  0.180                0.004                    97.8
 14                  0.219                0.024                    89.0
 15                  0.314                0.034                    89.2

Average Precipitation, % = 92.1%
    Absorbance as measured on the colorimeter; Absorbance being directly
      proportional to the amount of 'iron present.
                                    157

-------
      Table  D-2.  IRON OXIDE DATA, WITH FOG ADDED TO THE PRECIPITATOR
                                          Corona
                                            On                 Precipitation
Run                   "A"                   "A"                      (%)

  1                  0.180                0.0                      100.0
  2                  0.292                0.0                      100.0
  3                  0.094                0.0                      100.0
  4                  0.513                0.0                      100.0
  5                  0.133                0.0                      100.0
  6                  0.064                0.001                     99.5
  7                  0.111                0.0                      100.0
  8                  0.105                0.0                      100.0

Average Precipitation, % = 100'. 0
    Absorbance as measured on the colorimeter; Absorbance being directly
      proportional to the amount of iron present.
                                     158

-------
Potential advantages of charged droplet systems in comparison to con-
ventional electrostatic precipitators include the ability to clean gases
at lower voltage levels and elimination of problems of dust reentrain-
ment and of dust resistivity frequently encountered in conventional
electrostatic precipitators.  Advantages with regard to wet scrubbers
include lower water usage and lower pressure drop at the same collection
efficiency.

Disadvantages in comparison with conventional electrostatic precipitators
include:

1.  Need to produce and charge drops;

2.  Premature space charge precipitation of drops on walls of device;

3.  Neutralization of charge of drop upon collection of oppositely
charged particulates; and

4.  Need to handle wastewater stream.

Additional research and development activity on charged droplet systems
is recommended.  Emphasis should be placed on experimental work on bench-
scale or pilot-scale model systems so that meaningful engineering design
data are obtained.
                                   159

-------
                             REFERENCES

 1.  Kostow, L. P., "Design and Testing of Space-Charge Frecipitators,"
       M.S. Thesis, University of California, Berkeley, California,
       6 March 1972.

 2.  Melcher, J. R., and K. S. Sachar, "Charged Droplet Technology for
       Removal of Particulates from Industrial Gases," Environmental
       Protection Agency, Contract No. 68-002-0018, Massachusetts Insti-
       tute of Technology, Cambridge, Massachusetts (1971).

 3.  Private Communication, Dr. Michael Pilat, University of Washington,
       September 1973.

 4.  Private Communication, M. B. Koppang, TRW, Inc., September 1973.

 5.  Penney, G. W., "Electrified Liquid Spray Dust Precipitator," U.S.
       Patent No. 2,357,354 (1944).

 6.  Kraemer, H. F., and H. F. Johnstone, "Collection of Aerosol Particles
       in Presence of Electrostatic Fields," I.B.C. 47 (12), 2426-2434,
       (1955).

 7.  Melcher, J. R., "Progress Report—Research on Systems of Charged
       Droplets and Electric Fields for the Removal of Submicron Particu-
       lates from Industrial Gases," MIT Report, EPA Contract No. 68-002-
       0018, April 1973.

 8.  Faith, L. E., S. N. Bustany, D. N. Hanson, and C. R. Wilke, Ind. Engr.
       Chem.. Fund.. 6, 519 (1967).

 9.  Webber, M. E., "Experimental Studies on Space-Charge Precipitation,"
       M.S. Thesis, University of California, Berkeley, California,
       5 September 1969.

10.  Private Communication, Dr. J. R. Melcher, MIT, September 1973.
                                    160

-------
                             APPENDIX E

           FOAM SCRUBBERS AS PARTICULATE CONTROL DEVICES

INTRODUCTION

The addition of surface-active agents to water produces a foam and the
effectiveness of the collection of particulates by bubbles in wet scrub-
bers can sometimes be increased by generating a foam.  Wet scrubbers of
this type, called foam scrubbers, have been reported as being effective
for the collection of fine particles.

A foam scrubber consists of at least two components:  the foam generator
and the particle collector.  If the foam does not collapse after the
particle collection step, a third component is required—a foam destroyer.
The potential of foam scrubbing to control particulates, especially fine
particles, is discussed in the following sections.  Available theoreti-
cal and experimental information is presented and assessed.

THEORETICAL ASPECTS OF FOAM SCRUBBING

Particulate collection in a foam scrubber is effected by the use of bub-
bles.  At present, a comprehensive theory for foam scrubbers is not avail-
able.  Some insight into the potential of the method can be obtained from
the analysis of particle collection by bubbles reported by Fuchs—' and
Calvert.2/  The analysis of Fuchs which pertains to the absorption of
aerosols by bubbles is general and does not relate to any specific equip-
ment configuration.  Calvert's analysis relates to sieve plate columns.
The more general analysis of Fuchs will be used to indicate the potential
performance of a foam scrubber.

Fuchsi/ has derived expressions for the ratio of the number of particles
deposited per centimeter of path to the total number in the bubble for
deposition by impaction, sedimentation, and diffusion.  This ratio is
termed the coefficient of absorption (or), and has the units of reciprocal
centimeters.
                                    161

-------
For inertial deposition,





                           9 vb T


                     011 "  2 Rb2




where vfa - Ascending velocity of bubble  (cm/sec)



      Rb = Radius of bubble  (cm)



and,



       T = Force of internal friction per square centimeter
           18ug
      D  = Diameter of particle  (cm)




      p. = Density of particle




      u_ = Viscosity of gas in poise.
       O



Substitution for  T  in Eq. (E-l) results in
                         9 Vb Dp Pp   1 Vb Dp Pp
For absorption due to sedimentation,
where  R^,,  v^  and  T  represent parameters described under Eq.  (E-l)

and g = acceleration due to gravity (cm/sec^).
                                    162

-------
By substituting the expression for  T  in terms of particle size, Eq.
(E-3) reduces to


                          3 g Dp  Pp    1  gDpPp
                   ».  __     - - - f -   ^  mmi 	    ~	F
                          Vb 18 *g   24 Rbvb *g                  (E'4)
For diffusive absorption,
                          1.8    -^-r                             (E-5)
where  D = diffusivity of particles (cm2 /sec), and other parameters have
the same representation as those mentioned above.

The collection absorption due to all three mechanisms  (a?t)  can be ex-
pressed as


                         01 1 = "l + US + "D •                       (E-6)


It is apparent from the original derivation by Fuchs—  that the coef-
ficient of absorption  a,  is, in reality, the efficiency of collection,
inasmuch as it is a ratio of the actual number of particles deposited per
bubble to the theoretical number of particles that can be deposited per
bubble.  It is apparent that the efficiency is primarily a function of
particle size  (D) ,  bubble velocity  (v) ,  and bubble radius
The preceding analysis is undoubtedly greatly simplified compared to the
actual processes in a foam scrubber, but it serves to point out some of
processes that are operating in such a device.  Fuchs states that the
fact that absorption of dust in bubbling can sometimes be increased by
adding surface-active materials to the water is partly explained by a
reduction in bubble size, partly by the increased collision efficiency
of dust particles with the walls of the bubbles, and partly by the com-
paratively large residence times of the aerosol in the foam cells.  Fuchs
also discusses the results of several studies on the influence of surface-
active agents on the collision efficiency of particles with bubbles and
droplets and concludes that the effect of wetability of particles becomes
significant only when the kinetic energy of collision is small.

                                   163

-------
EXPERIMENTAL STUDIES OF FOAM SCRUBBING

Limited experimental study of foam scrubbers has been reported in the
technical literature.  References 4-6 report good results for the absorp-
tion of aerosols by foams.  Reference 7 describes a Japanese patent for
a foam scrubber.  Experimental data on the system  were not reported.

Professor R. Mahalingham at Washington State University has recently
initiated a study of wet scrubbing using a high-expansion foam.—   No
experimental data are available from his work to date.

Javorsky—' describes a foam bed scrubber for separating fine dust from
gases.  Figure E-l presents a cut-away diagram of the system—the Centri-
foam Scrubber.  The Centrifoam Scrubber is a product of Alfa-Laval AB,
Separation Division, Tumba, Sweden.  The Centrifoam Scrubber as described
by Javorsky consists of a vertically arranged cylindro-conical vessel
with an inlet and distribution facilities for the contaminated gas at
the bottom, which is also fitted with a paraxial sludge outlet.  The
greater part of the vertical vessel is filled with a foam bed, which is
supplied continuously or at predetermined intervals from a foam inlet in
the cylindrical part of the vessel and is supported by the grid of the
gas distributor.

The foam, the chemical composition of which may vary considerably but is
generally quite simple, is dense yet mobile.  It may be obtained by dis-
persing air in a suitable liquid to such an extent that the liquid content
of the resultant foam will be 0.2-3% by volume.  The interfacial areas
enveloping the innumerable bubbles of the foam bed provide efficient bar-
riers to the gas-borne dust particles, which are intercepted.

In the Centrifoam Gas Scrubber, the particles are exposed to adhesion
forces that make them adhere to the moist bubble interfaces at even the
slightest contact.  Hence, in the Centrifoam Scrubber the size of the
dust particles is of minor importance, the smallest particles being
separated almost as efficiently as the bigger ones.

Javorsky reports that the Centrifoam Scrubber has been tested with a
variety of dusts and that collection efficiencies in the 90-97% range are
obtained.  Pressure drops are reported to vary from 6 to 28-mm water.

An installation on a foundry cupola in Sweden is also reported by Javorsky
to clean over 3,000 m^/hr of gas.   The amount of water required to make
the necessary foam does not exceed 20 liters/hr, the velocity of the gas
is 1.6 m/sec, the foam number (i.e., the gas/liquid ratio when making the
                                    164

-------
                           ELECTRODES FOR FOAM
                           BED LEVEL REGULATOR
                           OUTLET FOR
                           CLEANED GAS

                           FOAM BED

                           FOAM INLET


                           GRID
                           INLET FOR
                           CONTAMINATED
                           GAS
 SLUDGE OUTLET
Figure E-l.  The Centrifoam Scrubber—
                               8/
                165

-------
the foam) is in the vicinity of 60, and the counterpressure is about 30-ram
water at a bed thickness of 900 mm.  Operating costs were reported to be
about $0.10 per hour.

CONCLUSIONS

The limited information available suggests that the potential of foam
scrubbing to control fine particles is good if the residence time is
sufficient to collect fine particles.  More detailed information is re-
quired regarding the mechanism of bubble formation, splitting, and de-
struction.  Potential problems are formation of fine bubbles for good
collection, foam destruction, and secondary foaming.  Technical feasibility
and optimum performance will rely on understanding the mechanisms associ-
ated with these processes as they will affect total energy and residence
time requirements.
                                    166

-------
                              REFERENCES

1.  Fuchs, N. A., The Mechanics of Aerosols. Pergamon Press, New York
      (1964).

2.  Calvert, S., J. Goldshmid, D. Leith, and D. Mehta, Scrubber Handbook,
      APT, Inc., Contract No. CPA-70-95, July 1972.

3.  Stairmand, C. J., Trans. Inst. Chem. Engrs.. 28, 130, London  (1950).

4.  Luchnskii, G.. Zh. fix, khim.. L3, 302 (1939).

5.  Alkeseeva, B., M. Andronov, Lab. Prakt. U.S.S.R.. 1£(1), 18 (1941).

6.  Bronsiky, D., F. Diwoky, Chem. Met. Eng.. 47, 541 (1940).

7.  Mlura, Mitsugu, "Dust Collection Using Bubbles," Japanese Patent No.
      Sho 47-48988, December 1972.

8.  Javorsky, B. S., "Gas Cleaning with the Foam Scrubber," Filtration
      and Separation, March-April 1972.

9.  Private communication, Professor R. Mahalingham, Washington State
      University, January 1973.
                                    167

-------
                               APPENDIX F

           ELECTRIFIED FILTERS AS PARTICULATE CONTROL DEVICES

INTRODUCTION

One potential technique for obtaining improved particulate control devices
is to use electric fields in conjunction with the conventional fabric fil-
ter.  Particles are normally brought to the surface of a fiber by diffusion
and convection as well as by the effect of the flow field of the gas and
retained on the individual fibers comprising the filter by surface forces.
The static charges on most naturally occurring aerosols may also play a
role in the collection and retention of the particles.

The magnitude of electrical forces compared to other forces that can be
applied to particulates suggests that improvement in aerosol filtration
should occur when electric fields are present in conjunction with fabric
filters.  Theoretical and experimental studies of electrified filters are
discussed in the following sections.

INVESTIGATIONS OF ELECTRIFIED FILTERS

External fields, internal fields, and electrets have been used in conjunc-
tion with fiber filter systems.  Each of these techniques is reviewed
separately.

External Electrostatic Fields

If a filter is placed in an initially homogeneous electric field, the fibers
will be polarized to produce an inhomogeneous electric field near the fiber
surface.  Neutral particles entering the filter will be polarized by the
external field and will therefore interact with the field around the fiber.
Any net charge on the particles will also interact with the local field.

The problem of the deposition of aerosol particles from an air stream of
velocity vg  , in a homogeneous electrical field,  E , has been studied
theoretically by Zebel,i/ Kazutaka and Koichi,!/ and Rao et al.,1' among

                                      168

-------
others.  Zebel developed a general theory for deposition upon an isolated,
uncharged cylinder when the particles are charged or uncharged.  Rao et
extended the theory by including the effect of the closeness of fibers on
the deposition of charged particles by use of a three-cylinder model.  The
single-cylinder model of Zebel significantly overestimates the deposition
of particles in low porosity filters because of the neglect of the inter-
action of neighboring fibers.  The three-cylinder model proposed by Rao
et al. predicts that the deposition on a given fiber will decrease as the
porosity of the filter decreases—a result more in agreement with experi-
mental observations.

Several experimental studies of filters in external fields have been
carried out.  All report substantial enhancement of the filter efficiency.
Only KirschV has used a filter of known geometry.  Kirsch found that
T] = eg   where  c  and  y  are constants (y = 2/3, c = 1) for both
idealized and real filters and  T|  is the collection efficiency.

Internal Fields with Potential Applied to the Filter Elements

Walkenhorst^' has studied filters consisting of arrays of insulated wire
0.08 mm in diameter (0.5 mm apart).  Potentials of 600-1,000 V were ap-
plied to the wires, with the polarity reversed at frequent intervals to
reduce the effect of the polarization of the collected dusts.  Typical
results are shown in Figures F-l and F-2 for coal and quartz dusts at
different relative humidities and gas stream velocities.  The pressure
drop at 10 cm/sec was less than 0.001 in. of water, although measurements
were only extended to about 0.1 um particles, these results suggest that
efficient collection can be obtained for even smaller particles.  Electric
power consumption would be expected to be very low.

Electrets

Electrets are devices which have a "frozen-in" effective charge.  As a
consequence, electrets are the electrostatic analog of permanent magnets
since an electrical field is produced without the application of external
potentials.  This field can be stable for long periods of time.   If a
filter is constructed from filters of an electret, inhomogeneous electric
fields will occur around each fiber.   It is then possible to obtain
electrical enhancement of aerosol  filtration without applying external
fields or potentials.

Electrets have been investigated experimentally and theoretically by
Ziekman.-'   Filters were successfully formed from electret fibers which
                                    169

-------
•k
c

8
0.
Q
CO
"c
u
fll
a!



80

60

40

20

0
-*""*"r" o-^fi^^
Q^^A^^^^k^
~o" *'
-/'
-P 	







- / Coal Dust
-/ 0) °
i (2) •
(3) o

(4M
i i i
10 cm/sec
1 0 cm/sec
1 0 cm/sec
Field Not
80 cm/sec
i
RH 34%
RH 95%
RH 95%,
Reversed
RH45%
i i
            1.0   2.0   3.0   4.0

                Particle Diameter,
                                     5.0   6.0
     Figure F-l.  Collection efficiency of wire grid filters

                               with coal dust
    lOOr
    80
I  60
    40
&

 0)
 O)
 o
§   20

£
       Quartz Dust

(l)o 10 cm/sec RH 43%

(2) • 20 cm/sec RH 57-70%

(3) o 80 cm/sec RH 50 - 70%
            1.0   2.0    3.0   4.0


                Particle Diameter,
                                     5.0   6.0
   Figure F-2.  Collection efficiency of wire grid  filters

                             with quartz dust
                           170

-------
were Initially approximately 23 um in diameter.  These filters showed a
penetration of 0.72 micron di-octylphtalate aerosols of less than 1%
(compared to 80-907. for comparable "unelectrified" filters) at a pressure
drop of less than 0.04 in. of water and a flow velocity of ~ 1.25 ft/sec.
However, the penetration of the filter increases with participate loading
as shown in Figure F-3.  This is the result of the cancellation of the
electric field of the electret by the collected polarized aerosol par-
ticles.  The data shown in Figure F-3 were taken with a filter 3 cm in
diameter.  Consequently, for a maximum penetration of 1% at a particulate
loading of 2.3 g/m3 (1 grain/ft3), only about 0.006 m3 (0.2 ft3) of gas per
square foot of filter could be treated before regeneration or replacement
of the filter.  Since it is not clear how readily an electret filter can
be regenerated, this could be a major problem in an industrial applica-
tion.

CONCLUSIONS

Available theoretical and experimental information indicates that elec-
tric fields can improve the collection efficiency of fabric filters.  How-
ever, the lack of adequate data on filter systems suitable for industrial
applications precludes a firm judgment on the potential usefulness of an
electrified filter for industrial applications.  Some intuitive state-
ments of advantages and disadvantages can be proposed.

It may be possible to utilize a system composed of a conventional bag
filter with an electrification addition.  The field would be applied in-
termittently, e.g., just after cleaning, to maintain uniformly low penetra-
tion during the period of cake repair.  The same combination operated
continuously might have a positive benefit in reducing penetration of fine
particles.  The system might also combat re-entrainment of collected par-
ticles, however, deposited, due to electrostatic adhesion.

Another possibility involves usage of a more porous filter media in an
electrified filter system to increase capacity, operate at lower pres-
sure drop, or allow control of difficult sources.  The economics and
reliability of such a device is speculative.

Nearly all of the test work to date involved short-term effects.  The
test regimes were constructed to minimize or eliminate the other important
filtration mechanisms.  Further testing would be needed to document any
improvement in collection efficiency when these other mechanisms are in
operation, particularly with respect to fine particulates.
                                    171

-------
  10'
    rl
o
«
  10-2
   10'
r3
      10'
       r2
              '    '  i  i i  i i  i
i    i   i  i  i  i i tl
                                           I   Mil
                                                                            l!
    1C'1                    10U                     101
                       Aerosol Load,  mg
Figure F-3.   Penetration  load curve  for electret filter
'    '	'   I  l l  i il
                102

-------
Nearly all tests utilized extremely low concentrations of aerosol.  Con-
sequently, for a short test run the deposition on the filter was light.
Little information is available on the incremental effect due to elec-
trostatic benefication when a more normal filter cake is present.

No studies investigated cleaning mechanisms, possible difficulties in
cleaning, or penetration increase during cleaning.  Cleaning of an elec-
trified filter or an electrostatically assisted system will probably be
more difficult than merely reversing polarity or turning off the field.
There are some reservations as well when one considers introduction of
an electrified filter into industries which collect dust having an ex-
plosive or fire hazard.

In principle, enhanced performance of fabric filters appears possible by
the utilization of electric fields.  It is not clear whether the improved
performance of these filters will justify their use or whether electrical
filters will be adaptable to the control of emissions from industrial
processes.  Further analysis is recommended with emphasis placed on ex-
perimental investigations of fabric filter configurations actually utilized
in industrial applications.  Theoretical studies are not recommended be-
cause of the inherent limitations of theoretical models of fabric filtra-
tion.
                                    173

-------
                               REFERENCES

1.  Zebel, G., J. Colloid Science. 20, 522 (1965).

2.  Kazutaka, M. and linoya Koichi, "Estimation of Collection Efficiency
      of a Dielectric Fibrous Filter," Kagaku KSguku. 33^(7) 88-93 (1969).

3.  Rao, K. S. et al., "Collection of Dust by Fabric Filtration in an
      Electrostatic Field," Department of Mechanical and Aerospace
      Engineering, University of Notre Dame (EPA Grant No. AP-01303-01)
      (1973).

4.  Kirsch, A. A., Aerosol Science. 3, 25 (1972).

5.  Walkenhorst, W., Aerosol Science. 1, 225 (1970).

6.  Ziekman, P., "Aerosol Filtration by Electrified Fibrous Filter Mats.
      Ill," Report No. 1970-16, Chemisch Laboratorium RVO-TNO, Netherlands
      (1970).
                                    174

-------
                               APPENDIX G

           GRANULAR BED FILTERS AS PARTICULATE CONTROL DEVICES

INTRODUCTION

The ability of porous beds of granular materials such as sand, coke, sin-
tered materials, Raschig rings or similar materials to remove particles
from gas streams has been appreciated for many years.  Although granular
beds have found practical application in atomic energy facilities and
the filtering of small volume gas streams, application to industrial
sources of particulate pollution has been limited--especially in the
United States.  In spite of the lack of extensive study and utilization
of granular bed filters, enough evidence is available to suggest potential
areas or sources where granular bed filters might be used to advantage to
control fine particulates.

The following sections present a discussion of:  (1) theoretical studies
of granular bed filters; (2) experimental studies of granular bed fil-
ters; (3) areas of potential application; and (4) estimates of cost of
granular bed filters.

THEORETICAL STUDIES OF GRANULAR BED FILTERS

Filtration of aerosols in a granular bed filter may result from the
interaction of the following mechanisms:  Brownian diffusion, inertial
impingement, direct interception, gravity settling, thermophoresis, electro-
static attraction, and particle agglomeration.  The nature and size of the
dust or mist particles, the structure and size of the bed granules, and
bed operating conditions will determine which of the collection mechanisms
are dominant.  Theoretical analysis of the performance of granular beds is
difficult because of the many possible interactions of the mechanisms that
may contribute to aerosol collection in granular beds.

Attempts have been made to adapt the results of studies of aerosol fil-
tration by fibrous filters to granular bed filters.111/  Since the
mechanism of collection on granules is essentially the same as in fiber

                                    175

-------
beds or mats, it seems plausible that correlations developed for fiber
beds could be used to describe granular bed behavior.  Attempts to use
existing correlations for fiber beds and mats to predict the performance
of granular filters have met with limited success.  Extrapolations of
collection efficiency equations developed for fiber filter systems to
granular filters result in predictions of poor performance for granular
filters.  These predictions are generally at variance with experimental
observations.

The major problems appear to be:

1.  Size differences between the fibers used in typical filters (10 mn)
and the granules that are of greatest potential interest (6-60 mesh)
from the standpoint of pressure drop; and

2.  Lack of knowledge regarding the controlling size dimension for a
granular filter.

In granular filters, the key dimension may be the size of a small pro-
tuberance or angular sections on the surface of the granules and not an
average granule diameter.  Particle agglomeration occurring in the tor-
tuous paths between the collecting granules is probably also important
to the collection process.

At present, there are no useful models for aerosol filtration in granular
beds.  Our knowledge of the performance of granular beds has been obtained
essentially from experimental studies at both the laboratory and pilot-
scale level.  Pertinent results of experimental studies are presented in
the next section.

EXPERIMENTAL STUDIES OF GRANULAR BED FILTERS

Experimental investigations of aerosol filtration by granular beds have
been performed on fixed and moving beds and beds augmented by electro-
static and sonic forces.  Although the studies have not been extensive in
either category of granular beds, experimental work has identified many
of the salient features of granular bed filtration.

Fixed and Moving Beds

A major portion  of the initial experimental work on granular beds has
been done by investigators associated with the atomic energy field.
Limited work has been done on the filtration of aerosols emitted by vari-
ous industrial operations.  Reference 11 presents a review of published
information on the use of sand-bed filters in industrial and atomic
energy facilities.
                                    176

-------
Table G-l presents a summary of selected experimental studies of granular
bed filtration.  The studies summarized in Table G-l, while not encompass-
ing all the investigations conducted, are representative of the type of
studies performed to date.  The results of the experimental studies on a
laboratory scale indicate that granular bed filters perform in many re-
spects in a manner analogous to fiber filtration systems.  Specific per-
formance characteristics of granular filters identified by laboratory
studies include:

a.  At constant face velocity, there exists a particular aerosol size
that exhibits a maximum penetration of a granular bed.

b.  Percentage of penetration by an aerosol in a granular bed is a func-
tion of aerosol particle size, aerosol type, face velocity, bed depth,
degree of bed packing, granule particle size, and granule shape.

c.  Particle size of the aerosol corresponding to maximum penetration
decreases with increasing face velocity.

d.  Aerosol penetration decreases with increasing bed depth up to a
limiting bed depth.

e.  Maximum penetration by an aerosol occurs at lower face velocities
with increasing bed depth.

f.  Aerosol penetration decreases as the granule size is decreased.

g.  Aerosol penetration of granular beds composed of rough, irregular
granules is less than that of beds composed of smooth granules of the same
characteristic diameter.

h.  Physical characteristics of the aerosol influence bed performance—
aerosols which tend to agglomerate are more readily collected.

Meager results have been reported on pilot-scale studies of granular fil-
ters operating on industrial sources of particulate pollution.  Limited
data available indicate collection efficiencies in the range 85-98%.  Col-
lection efficiencies were found to increase with increasing grain loading-
indicating that agglomeration may contribute to particle collection in in-
dustrial applications.  Pressure drops in  the pilot-scale tests were re-
ported to be in the range 0.5-5.0 in. H20.
                                    177

-------
                                       Table C-l.  SUMMARY OF SELECTED EXPERIMENTAL INVESTIGATIONS OF GRANULAR BED FILTERS

I Laboratory Studies
a. Thomas and
Yoderi'















b. McFee and
Sedlet^




















(a) Bed Type: Fixed Bed
(b) Bed Material: Sand
(c) Average Sand Grain Diameters-
0.161 cm, 0.071 cm. 0.036
cm
(d) Bed Height: 3.6 and 7.6 cm











(a) Bed Type. Fixed Bed
(b) Bed Material: Sand
(c) Sand Grain Diameter: 0.036
to 0.071 cm
(d) Bed Height: 15.2 to
76.2 cm

















Aerosol: OOP
Aerosol Particle
Size: 0.1 - 1.0 urn
Face Velocity: 0.11
to 2.2 cm/sec












Aerosol: Pu-U-Mo
Alloy Fume
kerosol Particle
Size: Geometric
Mean •> 0.07 urn, Stan-
dard Deviation = 2.7,
(Discrete sizes ranged
from < 0.02 to 4 pa)
Face Velocity: 0.3 to
40. 5 m/min
Aerosol Grain Load-
Ing: Approximately
0.11 g/m3









Parameters Studied

Sand Particle Size, Sand
Particle Shape, Face
Velocity














Bed Height, Face
Velocity


















Bed Performance

a) Penetration varied
with sand particle
size -- decreasing
diameter of sand
b) Rough and Irregular
sand showed higher
collection effi-
ciency then smooth
sand
c) Collection efficien-
cies varied with
sand particle size.
face velocity, and
aerosol particle
size. Efficiencies
ranged from 40 to
99.8%
[a) Penetration through
15.2 cm of sand
to 0.57% over range
of face velocities
from 0.3 to 40.5
m/min . Maximum pene*
tratlon of 0.57%
occurred in range
of 12.2 to 24.4 m/min
(b) Penetration through
76.2 cm of sand
varied from 0.0041
at 8.5 m/min to
0.019% at 1.5 m/min
with increase in
bed depth up to
depth of 30.5 to 45. 7
cm, but relatively
small improvement
occurred for beds
of greater depth
Comments

Experimental result demon-
strated the existence of
an aerosol size of maxi-
mum penetration of about
0.3 urn. Particle size
of maximum penetration
decreased with increasing
face velocity.









Experimental results indi-
cate the percentage of
aerosol particle size.
face velocity, bed depth
and degree of packing of
bed. Experimental ob-
servations confirm and
extend findings of
Thomas and Yoder.











•vl
co

-------
Table C-l.  (Continued)
Invest liator










c. Kovach and
Hannani'








d . Paretskjr-






e . Zahradnlk ,
et. .iJ/







8/
f. Squires-






Granular Bed Configuration










(a) Bed Type: Fixed Bed
 991
b) Pressure drop: 1.3
cm H20



Contents











Penetration of fly ash
aerosol was leas than
that for OOP aerosol.
However, particle size of
fly ash varied over a
wide range, and effect
may be due to both
changes In particle size
and aerosol type .
Results In agreement with
those reported by other
Investigators.




Primary Intent of study was
to Investigate simulta-
neous removal of fly ash
and SO2 by shaft-filter
system.




High collection efficiency
may be due in part to
preagglomeration of fly
ash before entering
filter



-------
                                                                             Table G-l (Concluded)
Investigator
g . Ducon
Company^













II. Pilot Scale
Studies
a. Ducon g/
Company^








b. Strauss and
ThrlngiP./






Granular Bed Configuration
Bed Type: Fixed bed with re-
verse gas flow cleaning















Bed Type: Fixed bed with re-
verse gas flow cleaning
Bed Material Sand
Sand Particle Size: 760 urn






Bed Type • Crushed brick
Granule Size 0.25 to 0.76 cm
Bed Thickness: 2.5 to 26.7 cm





Test Conditions
Aerosol •
a) Iron oxide from
oxygen- lanced
electric arc
furnace
b) Iron oxide from
oxygen -lanced
open hearth
furnace
c) Nickel ore
d) Fly ash
e) Tale dust
f) Plastic dust
Face Velocity: 0.26
m/sec


Aerosol: Catalyst
dust from fluid
catalytic cracker
(primary cyclone)
Gas Flow Rate: 13.5
to 28.3 nrVmin
Gas Temperature : 700
to 900" F
Inlet Grain Loading:.
0.34 to 1.95 g/m3
Aerosol: Fume from
open hearth furnace
Face Velocity: 0.3 to
1.02 m/sec
Inlet Gas Temperature:
450° to 970"C
Inlet Aerosol Load-
Ing: 0.46 to 22.9
Parameters Studied
Aerosol Type, Inlet Grain
Loading (4.6 to 11.4
g/-3)
























Face Velocity, Bed Thick-
ness, Inlet Gas Temper-
ature, Inlet Grain
Loading




Bed Performance
a) Collection efficiency
ranged from 98 to
99.91
b) Pressure drop. 10.1 to
15.2 cm H20












a) Collection efficiency
ranged from 85 to
981
b) Bed performance In-
fluenced by face
velocity, back-
cleaning cycle, bed
particle size


a) Collection efficiency
ranged from 85 to
981
b) Pressure drop: 1.01 to
11.9 cm H20



Comment s
Collection efficiency was
found Co be higher for
the finer sized iron
oxide aerosols than for
coarse fly ash. Indi-
cates that physical
characteristics of aero-
sol are an Important
factor in performance of
filter -- possibly with
regard to agglomeration
behavior .





Collection efficiency of
bed varied with changes
in inlet grain loading --
increasing with increas-
ing grain loading





Variables not well con-
trolled. However, on the
average, higher collec-
tion efficiencies were
noted at higher inlet
dust losdings


co
o

-------
 Augmented  Beds

 The  collection  efficiency  of  granular beds can be increased by  the use of
 augmenting forces.   Limited work has been done with electrostatically
 augmented  granular  beds.   Experimental and theoretical work is  discussed
 in the  following  paragraphs.

The filtration efficiency of granular beds can be enhanced by electro-
static augmentation.  If the filtration region is immersed in an electro-
static field, the dust particles will be driven in a direction that tends
to increase the probability of impact between the particles and the fil-
ter medium.—'

If the dust particles are initially uncharged, the electric field will
induce an electrical dipole in each particle which tends to align each
particle with the electric field.   This tends to orient the particle in
such a way that the negative end of the dipole contacts the positive sur-
face charge of the filter medium,  and similarly,  the positive ends tend to
contact the negative surface charge of the medium.  The result of this is
that the collection efficiency is increased somewhat by the attractive
force between the dust particles and the filter medium.

The filtration efficiency of packed bed filters can be further enhanced
if the dust particles are electrically charged prior to introduction into
the bed.  If charged particles are introduced into the filter medium with
a superimposed electric field, the dust particles are driven in the direc-
tion of the electric field.  Thus, the probability of collection and the
retentive forces for the dust particles are both increased by the addition
of the electrostatic augmentation.

Particle collection theory for turbulent flow electrostatic precipitation
shows that particles are captured only in the boundary layer of the gas
stream adjacent to the collection electrodes.  In conventional electro-
static precipitators, the interelectrode space is largely inactive for
particle collection.  This problem is significantly reduced in the aug-
mented bed collector since the entire surface of the filter medium acts
as the collection surface.  This results in a significant increase in the
collection area for a given volume of collector in comparison to that for
a conventional precipitator.  It seems probable from the above considera-
tions that high collection efficiency, even for small particles, can be
attained with electrostatic augmentation.!!'

A study of the collection of submicron aerosol particles on electrified
granular beds was carried out by Research-Cottrell,  Inc., Bound Brook,

                                    181

-------
New Jersey.!!/  Collection of particles occurred in a bed of 3-6 mm glass
beads held between two screens maintained at different electrical poten-
tials (Figure G-l).  Upstream of the glass spheres was a zone where the
aerosol particles were given an electrical charge.  Data taken on collec-
tion of a 0.6-0.7 um methylene blue aerosol are shown in Figure G-2..3/
An efficiency of 99.5% was obtained In a bed under 5.1 cm thickness with
a pressure drop of 0.51 cm  wg and at the relatively high face velocity
of 0.6 m/sec (2 ft/sec).

POTENTIAL AREAS OF APPLICATION OF GRANULAR BEDS

The most viable applications of granular beds as devices to control fine
particle emissions are on sources which require the control of effluents
under both corrosive and high temperature conditions.  Metallurgical
operations of various types are potential sources where granular beds may
function as control devices.  A particularly attractive application would
be where the granular bed material can serve as charge material to a sub-
sequent processing step.

Granular beds have been used in Europe in recent years on selected sources.
While the use of granular beds in Europe has not been directed to the con-
trol of fine particulates, granular beds have been used successfully on
cement and lime kilns, asphalt dryers, and clinker-coolers.IV  Rex
Chainbelt, Inc., has introduced a dual-bed granular bed system into the
U.S. by becoming the U.S. licensee for Gesellschaft fur Entstaubungsanlagen's
gravel-bed filter system, which has reported collection efficiencies of
99.5% or better.  The Rex Chainbelt gravel-bed filters remove heavier dust
particles by expanding and slowing the velocity of the air volume—the
semi-cleaned gas is then passed over a bed of uniform-size  round-quartz
gravel, which breaks up the gas flow into many streams and allows the fine
particles to be trapped.  Detailed testing of this system has not been
conducted to date, and its performance capability is not established.

ESTIMATES OF COSTS OF GRANULAR BEDS

Meager quantitative data are available on the costs of granular filters.
Lapple, in a review of dust collectors, stated that fine-packed beds
(< 4 mesh) are suitable for particle diameters of < 1 um and have a
2-25 cm water pressure drop, a power comsumption of 0.007-0.07 kw/m^/min
(0.2-2.0 kw/1,000 cfm) and an initial cost of $71-$177/m3/min ($2-5/cfm);
coarse beds (> 4 mesh) can filter particles as small as 5 um diameter
with a pressure drop of 0.51-5.1 cm water and a power consumption of 0.0035-
0.018 kw/m3/min (0.1-0.5 kw/1,000 cfm).
                                     182

-------

oo
X                                                                         Electrified
                                                                         Beads
                                                   -V
                                              \       V
                                       Charging Zone   "="
                                  Figure G-l.  Electrified packed bed

-------
      50
      60

      70


      80



1    90
£.
 B>
 o
      95
 o
.*
'o
E
LU

2
 •  99.0
_JC
 O
£
    99.5
    99.9
             Packed Bed Performance Data on Methylene Blue Aerosol (0.5-0.7 /idiom.)
             1.   Particle charging equivalent to 2.44 m/sec (8.0 ft/sec)
                  Gas velocity in single wire duct-type charger
             2.   Figures in parenthesis ( ) denote pressure drops across bed
             3.   Each plotted point the average of 3 to 7 tests
                                Jl-70)
                                                         (2.35)
                                                                               (3.25)
                Note
                Efficiencies do not include charging section collection of approx. 50%
                   I
                              I
I
I
I
I
       0.4        0.6        0.8        1.0         1.2         1.4        1.6
                       Packed Bed Thickness, 6.0 Millimeter Spheres (inches)
                                                                                     1.8
    Figure G-2.   Performance  characteristics of  electrostatically augmented
                                     packed bed
                                             184

-------
Cost estimates of various granular bed filter configurations are reported
in Ref. 3.  Installed costs for new power plants, based on the assumption
of installation by the power company, for the granular beds functioning
only as dust collectors were estimated at $78-$102 m^/min ($2.20-$2.90/cfm),
By comparison, electrostatic precipitators were estimated to cost between
$67-$78/m3/rain ($1.90-$2.20/cfm) for the same installation.

CONCLUSIONS

With the possible exception of electrically augmented systems, granular
beds do not offer any significant improvements in fine particulate con-
trol.  Additional research on unaugmented systems is not recommended.
Electrically augmented systems should be evaluated in conjunction with
other electrified filter systems.
                                    185

-------
                               REFERENCES

1.  Dorman, R. C., "The Role of Diffusion, Interception, and Inertia in
      the Filtration of Airborne Particles," in Aerodynamic Capture of
      Particles. E. G. Richardson, Editor, Pergamon Press, London (1960).

2.  Paretsky, L. C., "Filtration of Aerosols by Granular Beds,11 Ph.D.
      Thesis, City University of New York (1972).

3.  "Evaluation of Granular Bed Devices," Avco Applied Technology Division,
      Avco, Inc., AVATD-0107-69-RR, June 1969.

4.  Thomas, J. W., and R. E. Yoder, "Aerosol Size for Maximum Penetration
      Through Fiberglass and Sand Filters," A.M.A. Archives of Industrial
      Health. 1£, 545 (1956).

5.  McFee, D. R., and J. Sedlet, "Plutonium-Uranium-Molybdenum Fume Char-
      acteristics and Sand Filtration," Journal of Nuclear Energy. 22,
      641-650 (1968).

6.  Kovach, J. L., and D. G. Hannan, "Combined Dust, Aerosol and Vapor
      Removal Efficiency of Packed Activated Carbon Beds," Staub-Reinhalt
      Luft. 30_ (12),35-39 (1970) (English Translation).

7.  Zahradnik, R. L. et al., "Simultaneous Removal of Fly Ash and S02
      from Gas Streams by a Shaft-Filter-Sorber," Environmental Science
      and Technology. 4 (8) 663-667 (1970).

8.  Squires, A. M., and R. Pfeffler, "Panel Bed Filters for Simultaneous
      Removal of Fly Ash and 802:  I.  Introduction," Journal of APCA.
      20, 523 (1970).

9.  Kalen, B., and F. A. Zeng, "Filtering Effluent from a Cat Cracker,"
      Chemical Engineering Progress. 69 (6) 67 (1973).
                                    186

-------
10.  Strauss, W., and M. W. Thring, "Studies in High Temperature Gas Col-
       lecting," JournaJ^of_the_jTOn_and_S^^             191, 62 (1960).

11.  Juvinal, R. A. et al., "Sand-Bed Filtration of Aerosols:   A Review of
       Published Information on Their Use in Industrial and Atomic Energy
       Facilities," Argonne National Laboratory Report, ANL-7683, June 1970.

12.  Oglesby, S., and G. B. Nichols, "A Manual of Electrostatic Frecipitator
       Technology, Fart I - Fundamentals," Southern Research Institute,
       August 1970.

13.  Cole, W. H., U.S. Patent No. 2,990,912, "Electrical Precipitation and
       Charged Particle Collecting Structure," 4 July 1961.

14.  Englebrecht, H. L., "The Gravel Bed Filter—a New Approach to Gas
       Cleaning," Journal of APCA. 1£ (2) 43 (1965).

15.  Lapple, C. E., "Processes Use Many Collector Types," Chemical
       Engineering. 5£ (5), 144 (1951).
                                    187

-------
                              APPENDIX H

                       FLUID BEDS AS PARTICULATE
                            CONTROL DEVICES
INTRODUCTION
Investigations of fluid beds as particulate filtration media are quite
limited.  Only a few studies have been made over the years even though
fluidized beds offer advantages of low pressure losses and ease of
replacement of the bed for recycling during operation.  Further, porous
granular materials are especially suited for liquid type aerosols or to
the presence of high humidities where fiber filters clog rapidly.  The
major experimental studies of fluidized beds for particulate filtration
are reviewed in the next section.

EXPERIMENTAL STUDIES OF FLUID BEDS

Conventional Fluidized Beds

References 1-10 present the major studies that have been performed on
aerosol filtration by conventional fluidized beds.

Meissner and Mickleyl' conducted laboratory studies of sulfuric acid mist
removal via fluidized beds.  Exploratory runs on the removal of ammonium
nitrate dust were also conducted.  The experimental studies covered the
collection of sulfuric acid mist droplets, 2-14um in diameter, with fluid
beds of glass beads, silica (both nonporous), alumina and silica gel
(both porous), and aluminum silicate.   Using a 4.7 cm diameter glass
column,  1.2 m high, these  investigators found that for particles in the
micron range (~8 pm) an empirical equation of the form

                           In £i = -Kvgwn                       (H-l) (1)


could be fitted to their limited data.  In Eq. (1)
                                     188

-------
                   C, = inlet concentration

                   G£ = outlet concentration

                    K = 0.142

                   v  = superficial velocity (ft/sec)
                    O

                    w = bed weight/unit area (lb/ft2)

                    n = empirical constant which varied from 0.157 to
                          0.34 depending upon type of bed material.

Collection efficiencies up to 93% were obtained in the tests, and the
authors stated that with optimum design, higher efficiency might be
attained.

Scott and Guthrie-' conducted studies with f luidized beds using dioctyl
phthalate droplets (0.5-1.1 urn).  The average droplet diameter in all
runs was 0.87 urn.  Scott and Guthrie found that removal efficiencies
were only slightly affected by a 50% change in inlet mist concentration
or by variations in the weight of bed solids per unit cross sectional
area of bed.  Collection efficiencies varied from 70% at high superficial
velocities to 90% at low velocities.  Scott and Guthrie reported that
their data could be correlated by an equation of the form


                             In -^ = Kv-°'78                       (H-2)
                                L>2     6

In the studies reported in Refs. 1 and 2, it was found that the collection
efficiency was independent of the inlet concentration.  For the larger
droplets (2 to 14 microns) inertial impaction appeared to be the predomi-
nant collection mechanism, efficiency improving with increasing superfi-
cial gas velocity.  For the smaller dioctyl phthalate droplets, diffusion
collection predominated, as was shown by the opposite trend of a decreas-
ing efficiency with increasing superficial gas velocity.

Black—' has conducted one of the more detailed studies of fluid beds as
particulate filters.  The purpose of his work was to investigate the
factors contributing to removal efficiencies of small diameter aerosols
in a bed of fluidized glass shot.  A column geometry was chosen for the
study.  Figure H-l presents a schematic of the apparatus used in the
investigation.
                                     189

-------
                  I  PhotometeT  [
 C     CT"
Vacuum  Pump
   Filter
W Flowmeter
      4-Way
      Reversing
      Valve
     NH^CI Generator    Stirred Settling Chamber
                        Fluidized
                        Column
  Figure H-l.  Schematic flow diagram of  Black's fluid bed system
                                  190

-------
In Black's experimental program, the effectiveness of a  fluidized bed  in
removing airborne particulate from an air stream was investigated at
superficial gas velocities of 2.7 to 7.6 m/mtn.  Lower velocities were
insufficient to fluidize the bed material, while higher  velocities
resulted in excessive bed carry-over.  Bed height-to-diameter ratios were
varied from two to six.  The aerosols chosen were ammonium chloride and
tobacco smoke.  Concentration of aerosol ranged from 0.03 to 8.3 mg/cu m.
Ambient temperature conditions prevailed, normally 20 °C  to 30°C.

Light field microscopy was used to determine the inlet aerosol particle
size distribution for ammonium chloride, the mean size with respect to
count was 0.52 micron with a geometric deviation of 2.32.  Mean size with
respect to mass was 4.3 microns.  Tobacco smoke was determined by light
field microscopy to have a mean size of 0.35 micron.  However, it is felt
that this determination was too low since water and certain soluble
organics which surrounded the tobacco particle did not show up in the
method of microscopy.  Tobacco smoke particle diameter was estimated to
be the same as ammonium chloride particles, inasmuch as  filtration effi-
ciencies were almost exactly the same under the same conditions of operat-
ing variables.

Filtration efficiencies of the fluidized bed in removing either ammonium
chloride or tobacco particles of submicron size ranged from approximately
50 to 90% on a count basis.  Lowest efficiencies were encountered at
highest gas flow rates and lowest bed heights.  Highest efficiencies
resulted from low gas flow rates and high bed heights.  No effective
change in filtration efficiencies of the fluidized bed were found as a
result of bed age or changes in inlet aerosol concentration.  Black
reported that the following equation could fit the data  for all runs
within experimental accuracies:
                                         .0.4
                                 = °'565                            
where  f]  = collection efficiency, h is the bed height-to-diameter ratio,
and vg is the superficial gas velocity.  Over 90% of measured efficiency
values are within 5% of values predicted by Eq. H-(3).  Figure H-2 indi-
cates pressure drops for several different bed heights as a function of
gas flow rates.

The pressure drop across the filtration system does not change radically
with changes in gas velocity, and changes in gas velocity do not have
appreciable effect on filtration efficiencies.  This is not true with
many filtration systems.
                                     191

-------
VO
                      o  "F
                      CM
K

2
u


Q.

s
a

£
                                                     -o-o-o
                                                                           = 2
                                                                       '   i  i   i  i i  i 1
                                                        10
                                                                                     100
                                         Superficial Gas Flow Rate,  v  (ft/min)
                           Figure H-2.  Results of  pressure drop measurements across

                                          fluidized bed at various gas flow rates


                                          and bed heights

-------
Analysis of possible filtration mechanisms that might be effective in his
system, led Black to conclude that interception, supplemented by Brownian
diffusion, was the dominant mechanism.  Using variables suggested by
Fr iedlander ' s equation for effective filter efficiency with combined
effects of interception and Brownian diffusion gives
                                                                    (H-4)
Data for the various bed height-to-diameter values (h) used in the study
were plotted as shown in Figure H-3.  The curves in Figure H-3 can be
represented by Eq.  (H-5).

                    /„  1/3™  1/6M  1.78,M0.4S , „ 104
                 =  (NPe   NRe   %)     (h)   5.1 x 10                   ^
The slope of the curve for pure diffusion  (Equation H-4) is one while the
slope of the curve for pure interception is three.  The fact that the
slope of the lines in Figure H-3 is between one and  three implies both
Brownian diffusive and interceptive forces are active in the fluidized
bed used by Black.

The experimental conditions selected in the studies reported in Refs. 1,
2, and 4 emphasized different collection mechanisms.  The major differ-
ences between the work of Meissner and Mickley and that of Black or Scott
and Guthrie is that the average participate mass used in the work of the
latter investigators was two to three orders of magnitude smaller than
that used by Meissner and Mickley.  As a result, inertial forces were
more important in the experiments of Meissner and Mickley, while diffusion
and/or interception were the controlling mechanisms in the other experi-
mental configurations.

Jackson and coworkers have recently reported results of a study to compare
directly the collection of fine particles by a bed of granules operating
in both the fixed and fluidized states.   Monodispersed aerosols of di-
octyl phthalate, in sizes of 0.67 and 1.2 microns were collected in a bed
of granules of porous activated alumina having a mean size of 175 microns.
Collection efficiency was determined by simultaneous measurements of laser
                                      193

-------
     10
     9
     8
     7
     6
      1.
                                         oh = 2
Figure H-3.  Collection efficiencies in fluidized beds
                               194

-------
 light  attenuation by  inlet  and  outlet  streams.  A nuclei  particle counter
 confirmed results.  Bed depths  of  2.5-10.1. cm resulted in collections up to
 and  exceeding  997. in  the  fixed  state for either particle  size;  efficiencies
 dropped markedly upon bed f luidization and with increasing gas  velocity to
 70-807. at twice the initial f luidization. velocity.   The data  of Yankel,
 smoothed and adjusted for the region approaching  100%,  are shown in Figure
 H-4.-' The bed held 77. DOP  before  freezing.

 Reference 7 discusses some  alternative bed  designs which  permit periodic
 collapse of bubbles and mixing  to  increase  collection.  Arrangements to
 provide multiple contact  of gas flow and collecting  granules  include use
 of either multiple beds (stacked)  in 'series or perforated plates or other
 devices within a single bed.

 Attainment of  high efficiencies by staging may be possible even though
 single stage efficiencies are not 'high.  Assuming complete mixing between
 stages, the collection by "n" stages would give an outlet concentration
 of particles of

                        C2  = G! (1 - E')n= GI Pn                    (H-6)


where  E1 is the fractional  efficiency  per stage, P =  (1 - E1) is  the  pene-
 tration per stage, and GI,  C2 represent the inlet and outlet  concentrations.
 The  overall collection efficiency  is:
                      % = 100  (GI - C2)/C1 = 100 (1 - Pn)            (H-7)

For various combinations of stage efficiency, number of stages, bed depths,
and for a gas velocity (20 cm/sec) twice the initial fluidization velocity
(see Figure fl-4  and 0.67 micron size aerosol, the total collection effi-
ciencies are given in Table H-l.  Because the pressure loss across a
fluidized bed is less in inches of water than the depth for most materials,
(n x bed depth) gives a measure of the energy consumption.  From Table H-l,
the very shallow 2.5 cm bed at five stages, and less than 13 cm of water
pressure drop, gives the same very high collection as four stages of 5.1
cm bed and three stages of a 7.6 cm 'bed.  Other pressure losses may occur
between stages and an optimization analysis between collection efficiency
and energy loss is indicated.
                                     195

-------
    100
    95
    90
u
Qi
U

z
LU

u
Z
o

u
o
u
85
80
     75
     70
     65
                       Fixed-
                           •Fluidized
                      I
                                                1
                0.25            0.50           0.75



                       AIR VELOCITY, FT/SEC


Figure H-4.  Collection of 0.67 urn DOP aerosol by beds of

                     175 urn alumina granules
                                                                   1.00
                                 196

-------
Table H-l.  ESTIMATED COLLECTION BY MULTIPLE-CONTACT FLUIDIZED
                2.5 cm (1 in.)        5.1 cm (2 in.)       7.6 cm (3 in.)
 Number         Bed (E1 = 0.72)       Bed (E1  = 0.77)      Bed (E1 = 0.85)
of Stages       	
    1                 72.0                  77.0               85.0
    2                 92.2                  94.7               97.8
    3                 97.8                  98.8               99.7
    4                 99.4                  99.7               99.9
    5                 99.8                  99.9               99.99
 Equation H-7 is valid providing that the collection efficiency on each
 stage does not vary with changes in aerosol concentration.  Available data
 reported in Refs. 2 and 4 indicate that collection efficiency is not de-
 pendent upon aerosol concentration.  Scott and Guthrie report only a
 slight influence of aerosol concentration on efficiency for changes in
 concentration up to 50%.l/  Black*/ reports no effect for concentrations
                           .      o
 ranging from 1 to 75 x 10"* gm/nr.

 The main observations from previous work on fluid beds are:

 1.  The efficiency of collection in a fluid bed is dependent upon the size
 of the particles to be collected, gas velocity, particle density, and size
 of the collecting surface.

 2.  Collection efficiencies in fluid beds are not affected significantly
 by variations in inlet grain loadings.

 3.  Single-pass fluidized beds of granules inherently show low collection
 efficiencies because of the passage of gas bubbles through the bed.

 4.  Fluidized beds may provide very high collection efficiencies when
 employed in a multistage arrangement or with an internal bed structure
 that collapses and reforms gas bubbles.

 5.  Pressure drops through the fluid beds used in previous experimental
 work ranged from about 1.0 to 20 cm H^O which is no greater than that
 through other devices of comparable efficiency.
                                      197

-------
6.  Large gas volumes can be handled, though linear bed velocities may
only range from 0.3-0.6 m/sec (1-2 ft/sec), by use of larger bed areas
such as in drying operations which use fluidized beds.

7.  Fluid beds have no specific limitation with respect to temperature,
and they could be operated continuously and the bed solids regenerated.

Electrostatic Augmented Fluidized Beds

The mechanisms involved in electrostatic filtration of aerosols in fixed
and fluidized granular beds were studied at the Air Cleaning Laboratory
                                        a I
at Harvard University from 1955 to 1958.-'   Polystyrene spheres were used
as the granular media in these studies.  Polystyrene granules were charged
in situ by means of interspersed wires in the filter matrix or were re-
motely charged using a vibrating cylindrical Lucite trough.  The test
aerosol of gentian violet microspheres was charged to 18-64 electron
charges (positive)/ particle by a spinning disc generator.  A fixed bed
of polystyrene granules (280 micron diameter) with a surface charge den-
sity of 0.09 esu/cm2 had a 647. collection efficiency for atmospheric
dust as compared with a 96% efficiency for a fluidized bed expanded to
1207. of the original bed depth.I/

Studies were also conducted on beds of heated and unheated polystyrene.
Polystyrene granules possess an intrinsic bipolar charge at zero net
levels, and an added net charge acquired by triboelectrification.  Both
charges increase filtration performance.' Heating for 2 hr at 80°C causes
the granules to reverse their charges; mixed beds of heated and unheated
materials have decided bipolar effects.  Calculations indicated that a
fluidized bed of heated and unheated polystyrene granules will filter
five times as efficiently (based on target efficiency) as a bed of grounded
conducting media, handling over two times as much air volume at equivalent
flow resistance.I/

Professor Melcher and his colleagues at MIT have recently proposed the use
of electrofluidized beds for the collection of submicron particulates as
an outgrowth of work on charged droplet scrubbing .IP./  Figure H-5 presents
a schematic diagram of an electrofluidized bed which might be used for
the collection of oil ash particulate.  The electrofluidized bed is com-
posed of oil ash particulate that has achieved the size range of
50-100 microns.  These particles exist in a fluidized state with the
stack gas providing the fluidization.  Electrodes are used to impose an
electric field that serves not to collect particles but rather to charge
the collection sites.
                                     198

-------
 \
  A
                                                    outlet of cleaned gas
electro-
fluid! zed
bed
                     A
                                    i	z
inlet of gas
to be cleaned
corona charging of fine particulates
 Figure H-5.  Possible configuration for electrofluidized bed, typical
                     of application to cleaning of  oil ashi2/
                                       199

-------
Start-up of the system would be accomplished by seeding the bed with
particles which are either externally "manufactured" or residual from
previous operation of the bed.  The fine particulate enters the fluidized
region after having been charged in a conventional corona charging sec-
tion.  On passing through the fluidized bed, agglomeration is electro-
statically induced with the large particles.  The large particles are
withdrawn from the bed at a relatively slow rate so as to insure that
the agglomeration will itself retain the required population of collec-
tion sites.  Thus the final form of collection is in relatively large
particulate (50-100 urn range) withdrawn in a fluidized state from the bed.

Although it is too early to assess its ultimate utility in the control
of particulates from industrial sources, the electrofluidized bed offers
an interesting approach to control of fine particulates which is worthy
of some additional investigation.

CONCLUSIONS

The capability of fluid beds to remove dusts and mists from gas streams
has been studied only to a limited extent and their potential as par-
ticulate filters has not been completely characterized.  Multistage sys-
tems might provide high collection efficiencies, but operational problems
involving bed regeneration and attrition are anticipated.  Additional
work on conventional fluid beds is not recommended.

Fluidized beds augmented by electrostatic forces appear to be more
effective for the removal of particulates than are conventional fluidized
beds.  Additional analysis of the feasibility of augmented fluid beds is
recommended in conjunction with a general detailed assessment of electri-
fied filters.
                                    200

-------
                               REFERENCES

1.  Meissner, H. P., and H. S. Mlckley, "Removal of Mists and Dusts from
      Air by Beds of Fluidized Solids," Industrial and Engineering Chemistry.
      U, 1238 (1949).

2.  Scott, D. S., and D. A. Guthrie, "Removal of a Mist in a Fluidized Bed,"
      Can. J. Chem. Engr.. 37, 200 (1959).

3.  Black, C. H., and R. W. Boubel,  "Effectiveness of a Fluidized Bed in
      Removing Submicron Farticulate from an Air Stream," I/EC Process
      Design and Development. £, 573 (1969).

4.  Black, C. H., "Effectiveness of a Fluidized Bed in Filtration of Air-
      borne Particulate of Submicron Size," Ph.D. Thesis No. 67-5646,
      University Microfilms, Ann Arbor, Michigan (1967).

5.  Boubel, R. W., and D. C. Junge,  "A Study of  Fluidized Bed Filters Sub-
      jected to High Radial Accelerations," presented at the 64th Annual
      Meeting, AIChE, San Francisco, California, November 1971.

6.  Yankel, A. J., R. G. Patterson,  and M. L. Jackson, "Fine Particle Col-
      lection with a Fixed-Fluidized Bed," under review by Industrial and
      Engineering Chemistry, Process Design and Development, 1973.
      Presented at the Annual Meeting, Air Pollution Control Association,
      Pacific Northwest International Section, Eugene, Oregon, 15 November
      1972.

7.  Jackson, M., "Fluidized Beds for Submicron Particle Collection,"
      presented at the 75th National Meeting, AIChE, Detroit, Michigan,
      3-6 June 1973.

8.  Juvinall, R. A. et al., "Sand-Bed Filtration of Aerosols:  A Review
      of Published Information on Their Use in Industrial and Atomic
      Energy Facilities," Argonne National Laboratory Report  ANL-7683,
      June 1970.
                                   201

-------
 9.  Charkravarty,  R.  K. et al., "Application of Fluidized Technique in
       Gas Purification," Indian Journal of Technology. 1, 423 (1963).

10.  Melcher, J. R., "Progress Report—Research on Systems of Charged
       Droplets and Electric Fields for the Removal of Sub-Micron Par-
       ticulates from Industrial Gases," MIT Report on EPA Contract No.
       68-002-0018, April 1973.
                                    202

-------
                               APPENDIX I

         GAMMA-RAY PRECIPITATORS AS PARTICULATE CONTROL DEVICES

INTRODUCTION

The idea of utilizing ionizing radiation to charge particulate matter so
that it could be collected by electric fields in an electrostatic pre-
                                                     -I  O/
cipitator has been proposed in the patent literature.•=!£'  Schmidt has
recently proposed the use of gamma-ray ionization to charge particles and
has proposed the development of particulate collection devices utilizing
gamma-ray ionization3/  Available theoretical and experimental informa-
tion on this concept is reviewed in the following sections.

THEORETICAL ASPECTS

In a radiation-induced precipitator, electrons and positive ions are pro-
duced throughout the volume of the conduit by gamma rays interacting with
the walls and gas (Figure 1-1).  Several processes are involved.  Photo-
emission can take place at the walls, freeing large numbers of electrons.
Direct ionization of the gas can take place; recombination of electrons
and positive ions can occur in this high pressure, low temperature plasma
system; the dust particles themselves can be ionized by direct gamma-ray
interaction, etc.  Electrons have a tendency to combine with polar and
electronegative gases present in the air, such as S02,  H20 and 02.  These
form negative ions, which are 1,000 times less mobile than electrons, and tend
to recombine with, and neutralize, the positive charges.  Space charge
separation combined with photoemission may occur.  These effects are
represented in Figure 1-2.

Photoemission of electrons from the anode, traveling against the field,
could contribute to the negative part of the space charge, while photo-
emission from the cathode,  in the direction of the field, could further
energize the electrons, thereby reducing their recombining potential in
the positive space charge section.
                                   203

-------


^-*~

(rJSLr^RECOM
X^y w m ^*^x

f*O ANEGAT
INATION
/E
V(M)« 	 VION FORMATION
- -^'"g
^-^)lONIZA
•*_ *i *~~^ _^_
XRAYS

© ©
FLUE GAS
! t
PHOTO-
EMISSION
ION
•x.^
e- -ELECTRON
+

\ /
" - +A+
NEGATIVE! © POSITIVE
SPACE SPACE
CHARGE o CHARGE



DIRECTION OF .
A FIELDI 1
t
GAS FLOW
         ©  -POSITIVE  ION

         ©  - NEGATIVE ION
         ®  - NEUTRAL  MOLECULE
        ©  - PARTICLE

Figure 1-1.   Reactions occurring in the
  radiation-induced precipitator
Figure 1-2.   Space  charge separation and
  particle charging in radiation-induced
  precipitator
                                         204

-------
Particles entering the ion-filled region become charged by field charging
and diffusion charging.  The charging processes are identical to those
that occur in a conventional electrostatic precipitator.  The charged
particles then would enter an electric field and be collected in the
usual manner of an electrostatic precipitator.

EXPERIMENTAL INVESTIGATIONS

Experimental studies of the gamma-ray precipitator concept have been con-
ducted at Penn State University using the experimental system shown in
Figure 1-3.^2.'  A small coal burning stoker-fed home-type furnace was
installed adjacent to the cobalt pool of the reactor facility.  The par-
ticulate and combustion products from the furnace travel upward to a
cast iron flue pipe assembly.  This assembly contained by pass regulating
dampers and sampling ports.  The sampling ports were used to extract gas
and particulate samples and to measure gas temperature.  The precipitator
section was a coaxial aluminum pipe section whose inner pipe was 12.7 cm
ID and whose outer pipe was 20.3 cm ID.  This section was insulated elec-
trically from the remainder of the system by a Teflon ring arrangement
at the top and three standoff insulators located halfway down the pre-
cipitator section .I/

Flue gas proceeded down the center pipe, turned around, and went up the
outer sleeve and, ultimately, to the blower and stack.  The gas  flow
rates used were in the range of commercially used value of 1.2-4.6 m/sec
(4-15 ft/sec).  Most of the experiments were run at 1.5 m/sec (5 ft/sec)

Irradiation occurred at the bottom turn around area of the precipitator
section.  The °®Co sources were in the form of pencils, doubly encapsulated
in stainless-steel tubes, 1.3 cm in diameter x 30.5 cm long.  The pencils
were arranged in a horseshoe-shaped holder on top of a movable platform.
The platform could be positioned to have the horseshoe encircle the large
aluminum pipe, or it could be moved away to reduce the field by given
amounts.  The shielding for the cobalt pencils was 4.3 m (14 ft) of water.—'

A 50,000-V, 5mA power supply was mounted on the scaffolding, and the
electrical connections were such that the inner pipe could be made either
positive or negative with respect to the outer pipe.  The outer pipe was
always grounded.—'

Figure 1-4 shows the overall efficiency of the device described in Figure
1-3 as a function of central electrode potential.  Efficiences were ob-
tained from gross filter samples taken at the inlet and outlet of the pre-
cipitator.  The efficiency at zero kV represents mechanical removal,
                                  205

-------
                    Building Wall

            I  f      . Blower
Dampers
Furnace
Shed
     Cast  Iron
     Assembly
                                                      Sampling  Ports
                                                       I   >/ Insulators
                                                                    I - Beam
                            lonization-
                            Precipitator
                            Chamber
                                    Standoff
                                    Insulators

                                    Lead  Jacket

                                    Cobalt  Source
                                                                           Platform
    Figure  1-3.   Schematic diagram of gamma-ray precipitator and auxiliary
                        equipment  at Pennsylvania State University
                                            206

-------
 mostly of larger-sized particles.  Error bars indicate data points where
 multiple samples were taken and standard deviations calculated.

 Figure 1-4 also indicates no change in efficiency when the polarity of
 the central electrode is changed.  The efficiency in both cases  for this
 single stage device is approximately 96% with an applied potential of
 27 kV across the precipitator tube.

 Figure 1-5 shows the importance of dose rate applied to the charging
 region of the precipitator for one run.  The maximum dose was used in all
 other experiments.

 Cascade impactor samples were used to study precipitator performance
 as a function of particle size distribution.  A typical incoming concen-
 tration of fly ash  from the furnace is shown in Table 1-1.  These values
 show a high proportion of small particles indicating a larger percentage
 of small particles than exists in most coal-burning commercial stacks.
Table 1-1.  TYPICAL PARTICLE CONCENTRATION AT INLET TO GAMMA-RAY PRECIPITATOR
                                (20-ft3 sample)

Mean
Diameter
(urn)
0.4 or smaller
1.1
1.8
2.7
3.8
5.4 or smaller



me/ft3
1.27
0.30
0.11
0.10
0.14
0.68
Percent
of Total
by
Weight
48.9
11.5
4.4
3.8
5.4
26.1


Number
(cm3)
6.8 x 103
7.7 x 103
6.6 x 102
1.7 x 102
8.7 x 10
1.5 x 102
Percent
of Total
by
Number
98.8
1.1
0.094
0.026
0.013
0.021
 Figure  1-6 presents  fractional efficiency data for the device.  The frac-
 tional  efficiency of the device is about the same as that for a high
 efficiency electrostatic precipitator.
                                    207

-------
£ 100
x 80
Q>
2 60
u.
UJ
I 4(*
8 20
3 o
^ 	 &AA-A*-
" /^'^
y
' • Central Electrode Negative
* Central Electrode Positive
-
i i I i i i i i
          4    8   12   16   20  24   28
          Central Electrode  Potential (kV)
                                      32
  Figure  1-4.   Collection efficiency of gamma-ray
                precipitator vs central electrode
                potential based on weight percent
                of  inlet particulate concentration
 u
 u
£
LLJ

J
"o
 9)
lOOr
 90
 80
 70
 60
1.0
               22-kV Applied Potential
               10.0
         Dose Rate
                                100.0
  Figure 1-5.  Collection efficiency of gamma-ray
                   precipitator vs dose rate
                        208

-------
               99.9
               99.0
            u
               90.0
                   0123456
                          Particle Diameter ( fi m)
Figure 1-6.  Experimental fractional efficiencies  of  gamma-ray precipitator
                                    209

-------
CONCLUSIONS

Definitive conclusions regarding the potential of the gamma precipitator
cannot be reached at this time.  Only limited test data on a nonrepresenta-
tive source are available.  However, the results of the limited tests in-
dicate collection efficiencies comparable to those conventional electro-
static precipitators.

The gamma ray precipitator has the following possible advantages in com-
parison to conventional electrostatic precipitators:  (a)  possible higher
ion densities; (b) lower operating costs; and (c) simplified cleaning.

The major disadvantages are the radioactive material required and the
shielding structures.  Both of these disadvantages will greatly limit
the use of such a device even if it were proven highly efficient.
Utility companies with experience in handling nuclear materials would
appear to be the only likely applications for such a device.

On balance, the disadvantages of a gamma ray precipitator far surpass the
advantages.  Development of this device is not warranted in our opinion.
                                    210

-------
                               REFERENCES

1.  McCray, H. F., U.S. Patent 1991934, "Apparatus and Process for
      Utilizing Emanations from Radioactive Material/1 September 1929.

2.  Jacob, C. W., U.S. Patent 2381455, "Electrical Precipitation Apparatus,"
      October 1942.

3.  Leupi, R. et al., U.S. Patent 2934648, "Apparatus for the Electrical
      Charging by Means of Radioactive Preparations of Matter Suspended
      in a Gas Stream," April 1955.

4.  Schmidt, C. M., Patent Application No. S.N. 83, 874, October 1970.

5.  Dickter, William and M. A. Schultz, "Investigation of a Device Using
      Radiation to Change and Collect Particulate Matter, Nucl. Techno1..
      12. 243 (1971).

6.  Schultz, M. A., M. E. Crotzer and W. R. Knapick, "Collection of
      Particulate Matter from Smokestacks Using Gamma-Ray lonization,"
      Nucl. Technol.. 15^ 38 (1973).

7.  Dictor, W., "An Investigation of a Device Using Radiation to Charge
      and Collect Particulate Matter," M.S. Thesis, Nuclear Engineering,
      The Pennsylvania State University, University Park, Pennsylvania
      (1969).
                                   211

-------
                                 TECHNICAL REPORT DATA
                          (Please read luUniclioiK on the reverse before completing)
 i RLPOOT NO  .
  EPA-650/2-74-027
                            2.
                                                        3 RECIPIENT'S ACCESSION-NO.
 J TITLE ANDSUBTITLE
  Control Technology for Fine Participate Emissions
             5 REPORT DATE
             May 1974
                                                        6. PERFORMING ORGANIZATION CODE
 7 AUTHOR(S)

  L.J. Shannon
             8. PERFORMING ORGANIZATION REPORT NO
9 PERFORMING OR9ANIZATION NAME AND ADDRESS
  Midwest Research Institute
  425 Volker Boulevard
  Kansas  City, Missouri 64110
             10. PROGRAM ELEMENT NO.

             1AB012;ROAP 1AB012-05
             11. CONTRACT/GRANT NO.
             Grant R-801615
 12 SPONSORING AGENCY NAME AND ADDRESS
  EPA, Office of Research and Development
  Air Pollution Control Division
  Washington, DC  20460
             13. TYPE OF REPORT AND PERIOD COVERED
              Final
             14 SPONSORING AGENCY CODE
 15 SUPPLEMENTARY NOTES
 16 ABSTRACT
  The report gives results of a study to assess the state-of-the-art of control
  technology for fine particulates emitted from stationary sources.  It emphasizes
  the analysis of control technology for primary particulates. The initial activity
  was a general review of collection and agglomeration mechanisms for particulates.
  Both theoretical and experimental facets were reviewed with attention on forces
  or collection mechanisms expected to be of importance for fine particles. The
  analysis of control technology for fine particulates centered on conventional control
  equipment, on emerging control technology, and on proposed or conceptual
  control systems.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                           b IDENTIFIERS/OPEN ENDED TERMS
                         C.  COSATI field/Group
  Air Pollution
  Dust
  Collection
  Agglomeration
 Air Pollution Control
 Stationary Sources
 Fine Particulates
 Primary Particulates
13B
11G
  DISTRIBUTION STATEMENT

  Unlimited
19 SECURITY CLASS (Tha Report)
 Unclassified
                         21
 NO OF PAGES

  224
20 SECURITY CLASS (Thispage)
 Unclassified
                         22. PRICE
EPA Form 2220-1 (9-73)
                                          212

-------