EPA-600/2-77-193
                                             September 1977
SECOND  EPA FINE  PARTICLE
    SCRUBBER  SYMPOSIUM
                       by

             Richard Parker and Seymour Calvert

               Air Pollution Technology, Inc
             4901 Morena Boulevard, Suite 402
                San Diego, California 92117
                Contract No 68-02-2190
                 ROAP No. 21ADL-029
               Program Element No. 1AB012
             EPA Project Officer: Dennis C. Drehmel

           Industrial Environmental Research Laboratory
            Office of Energy, Minerals, and Industry
             Research Triangle Park, N.C. 27711
                    Prepared for

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

-------
                          ABSTRACT

     The Second Fine Particle Scrubber Symposium was held on
May 2-3, 1977 in New Orleans, Louisiana.   The symposium was
sponsored by the U.S. EPA and was organized and hosted by
A.P.T., Inc.
     The conference was intended to stimulate and generate
new ideas for fine particle control using wet scrubbers, and
to promote the transfer of technology to  users.  The subject
matter was concerned with the collection  of fine particles
in any type of wet collector with emphasis on scrubber per-
formance data in industrial applications.
     This report presents the symposium proceedings, including
introductory remarks and the sixteen technical papers.
                             111

-------
                       ACKNOWLEDGEMENT

     A.P.T., Inc. wishes to express its appreciation for
excellent coordination and assistance in organizing and
conducting this symposium to Dr. Dennis Drehmel, EPA Pro-
ject Officer.
                             IV

-------
                      TABLE OF CONTENTS
                                                            Page
Abstract	iii
Acknowledgement  	   iv
List of Figures	   vi
List of Tables	xiv
Introductory Remarks	   1
Paper 1 - Wet  Scrubbing Experiences for Steel Mill
          Applications	   5
Paper 2 - Wet  Scrubbing Experience with Fine Borax Dust .  .  25
Paper 3 - F/C  Scrubber Demonstration on a Secondary
          Metals Recovery Furnace  	  35
Paper 4 - Fine Particulate Scrubbing - New Problems
          and  Solutions	61
Paper 5 - Entrainment Separators for Scrubbers	75
Paper 6 - Relationships of Collection Efficiency and Energy
          Dissipation in Particulate Scrubbers	97
Paper 7 - Diffusiophoretic Particle Collection under
          Turbulent Conditions	115
Paper 8 - Improved Design Method for F/C Scrubbing	141
Paper 9 - Interfacial Surface Effects on Particle
          Collection	163
Paper 10A- EPA Mobile Particulate Collectors -
          Program Intentions	187
Paper 10B- Cupola Foundry Particle Control with Venturi
          and Sieve Tray Scrubbers	195
Paper 11 - Field Test of a Venturi Scrubber in Russia. . .  . 209
Paper 12 - A.P.T. Field Evaluation of Fine Particle
          Scrubbers	221
Paper 13 - Results of Flue Gas Characterization Testing
          at the EPA Alkali Wet-Scrubbing Test Facility .  . 255
Paper 14 - The Waterloo Scrubber	279
Paper 15 - Fine Particle Control with UW Electrostatic
          Scrubber	303
Paper 16 - Fine Particle Scrubbing with Lone Star Steel
          Hydro-Sonic Cleaners "The Coalescer"	319
List of Attendees	339
                              v

-------
                       LIST OF FIGURES

Figure                                                      Page
                           PAPER 1
  1    No.  1 Boiler House, Middletown Works Scrubber
       System Flow Diagram	    21
  2    Sinter Plant, Houston Works Scrubber System
       Flow Diagram	    22
  3    No.  2 Open Hearth, Middletown Works Scrubber
       System Flow Diagram	    23
  4    Recycle Plant,  Middletown Works Scrubber
       System Flow Diagram	    24

                           PAPER 2
  1    Modified Scrubber	    33
  2    Emission Rate vs. Scrubber Pressure Drop
       for Triple Scrubber	    34
                           PAPER 5
  1    F/C Scrubber Demonstration Plant 	    51
  2    Flow Diagram of F/C Scrubbing System	    52
  3    Quencher Unit of F/C Demonstration Scrubber	    53
  4    Scrubber Unit of F/C Demonstration Plant 	    54
  5    Cooling Tower of F/C Demonstration Scrubber
       System	    55
  6    F/C Demonstration Scrubber Performance 	    56
  7    Particle Penetration Versus Aerodynamic
       Diameter for Run 64	    57
  8    Particle Penetration Versus Aerodynamic
       Diameter for Run 74	    58
  9    Particle Penetration Versus Aerodynamic
       Diameter for Run 69	    59

                           PAPER 4
  1    Competing Devices (Non-Scrubber)  	    67
  2    Comparison of New Scrubber Devices 	    68
  3    New Scrubber Applications	    69
  4    Manufacturers	    70
  5    Useful Addresses of Information Sources	    72
                              vi

-------
Figure                                                      Page
                           PAPER 5
  1    Tube Bank Configuration	   85
  2    Baffle Configuration 	   86
  3    Inclined Baffle Configuration	   87
  4    Experimental Penetration through Mesh,
       Upward Flow	   88
  5    Experimental Penetration through Packed
       Bed, Upward Flow	   89
  6    Experimental Penetration through Tube Bank,
       Upward Flow	   90
  7    Experimental Penetration through Horizontal
       Baffles, Upward Flow	   91
  8    Experimental Penetration through 30°  Inclined
       Baffles, Upward Flow	9Z
  9    Experimental Penetration through 45°  Inclined
       Baffles, Upward Flow	   93
 10    Entrainment Separator Performance Cut Diameters.  .  .   94

                           PAPER 6
  1    Flowsheet of Experimental Scrubber System	107
  2    Multiple-Orifice Series Gas/Liquid Contactor ....  107
  3    Spray Gas/Liquid Contactors	108
  4    Miniscrubber Signature Curves for Aerosols
       D,  E, F, and G	109
  5    Orifice Scrubber Performance Curve for  Aerosol  D  .  .  109
  6    Orifice Scrubber Performance Curve for  Aerosol  E  .  .  110
  7    Orifice Scrubber Performance Curve for  Aerosol  G  .  .  110
  8    Multiple-Orifice Series Scrubber Performance
       on  Aerosol D	Ill
  9    Multiple-Orifice Series Scrubber Performance
       on  Aerosol E	Ill
 10    Multiple-Orifice Series Scrubber Performance
       on  Aerosol G	112
 11    Performance of Spray Scrubbers on Aerosol D
       at  Values of f=0 - 0.39	112
 12    Performance of Spray Scrubbers on Aerosol D
       at  Values of f=0.90 - 0.99	113
 13    Performance of Spray Scrubbers of Ejector
       Configuration of Aerosol D	113
                             vii

-------
Figure                                                      Page

                           PAPER 7

  1    Schematic Diagram of Experimental Equipment	132

  2    Results for Helium-Ammonia Mixtures with 0.79
       Micron Diameter Particles	133

  3    Results for Methane-Ammonia Mixtures with 0.79
       Micron Diameter Particles	134

  4    Results for Nitrogen-Ammonia Mixtures with
       Particles of Various Diameters 	 135

  5    Results for Argon-Ammonia Mixtures with 0.79
       Micron Diameter Particles	136

  6    Results for Freon 12 (dichlorodifluoromethane)-
       Ammonia Mixtures with Particles of Various
       Diameters	137

  7    Results for Nitrogen-Trimethylamine Mixtures
       with Particles of Various Diameters	138


                           PAPER 8
  1    Generalized F/C Scrubber System	153

  2    Multiple Plate F/C Scrubber System ....  	 154
  3    Effect of Particle Diameter on Condensation Ratio. . 155
  4    Effect of Liquid Bulk Temperature on
       Condensation Ratio 	 156

  5    Effect of Gas Inlet Temperature on Condensation
       Ratio	157
  6    Effect of Particle Number Concentration on
       Condensation Ratio 	 158
  7    Effect of Liquid Heat Transfer Coefficient
       on Condensation Ratio	159
  8    Particle Size Distribution Before and After
       Condensation 	 160
  9    Particle Penetration vs Aerodynamic Dia.- Run 64  . . 161
 10    Particle Penetration vs Aerodynamic Dia.- Run 69  . . 162
                           PAPER 9

  1    Model for Collision of Particle and Liquid  Droplet  .  172

  2    Models for Surface Deformation 	  173

  3    High Speed Cine Microscope Equipment used for
       Experimental Evaluation of Coalescence Mechanisms.  .  174


                              viii

-------
Figure                                                      Page
                      PAPER 9  (Continued)
  4    Glass Rod with Simulated Fly Ash Particle
       Mounted on Traversing Mechanism	175
  5    Traversing Mechanism for Impacting Particles
       into Water Droplets.	176
  6    Coalescence of 2800 ym Diameter Glass Particle
       with Water Droplet, 1200 ysec Delay  (10,000 pps)  .  .   177
  7    Coalescence of 2800 ym Diameter Glass Particle
       with Water Droplet, 2067 ysec Delay	178
  8    Coalescence of 1700 ym Diameter Glass Particle
       with Water Droplet, 585 ysec Delay	   180
  9    Coalescence of 1000 ym Diameter Glass Particle
       with Water Droplet, 780 ysec Delay	181
 10    Coalescence of 725 ym Diameter Glass  Particle
       with Water Droplet, 468 ysec Delay	182
 11    Coalescence of 275 ym Diameter Glass  Particle
       with Water Droplet, 351 ysec Delay	183
 12    Coalescence of 100 ym Diameter Glass  Particle
       with Water Droplet, 156 ysec Delay	184
 13    Comparison of Theoretical  Predictions and
       Experimental Measurements  of Coalescence
       Delay Time	185

                          PAPER 10-A
  1    Scrubber Flow Schematic	192
  2    Plan View of Mobile ESP  Facility	193

                          PAPER 10-B
  1    Scrubber Flow Schematic	200
  2    Effect of L/G Ratio on Venturi Scrubber	201
  3    Effect of Throat Velocity on Venturi Scrubber.  .  .  .   202
  4    Effect of Pressure Drop  on Venturi Scrubber	203
  5    Effect of Presaturation  on Venturi Scrubber	204
  6    Fractional Removal Efficiency for Venturi
       Scrubber	205
  7    Effect of L/G Ratio on Sieve Tray Scrubber 	   206
  8    Fractional Efficiency of Sieve Tray Scrubber ....   207

                             ix

-------
Figure                                                      Page

                          PAPER 11

  1    NIKOPOL Scrubbing Apparatus	217
  2    NIKOPOL Scrubber Sampling Port Locations 	  218


                          PAPER 12

  1    Predicted and Experimental Penetrations for
       Koch Flexitray	237
  2    Predicted and Experimental Penetrations for
       Ducon Multivane Scrubber 	  238
  3    Predicted and Experimental Penetration for
       Mobile Bed Scrubber	239
  4    Predicted and Experimental Penetrations for
       Chemico Venturi	240
  5    Predicted and Experimental Penetrations for
       Encott Wetted Fiber Scrubber 	  241
  6    Predicted and Experimental Penetration for
       Sly Impinjet	242
  7    Predicted and Experimental Grade Penetration
       Curves for Venturi Rod Scrubber	243
  8    Predicted and Experimental Penetrations for
       AAF Kinepactor 32	244
  9    Predicted and Experimental Grade Penetration
       Curves for AAF Kinepactor 56	245
 10    Predicted and Experimental Penetrations for
       Variable Rod Scrubber	246
 11    Theoretical and Experimental Cut Diameters as a
       Function of Pressure Drop for Several Scrubber
       Types	247
 12    Efficiency of Single Drop Versus Inertia Para-
       meter at NRec} = 9.6 with Npj) as Parameter	248
 13    Experimental and Predicted Particle Penetration
       Versus Particle Diameter 	  249
 14    Experimental Grade Penetration Curve for CHEAP .  .  .  250
 15    Cut Diameter Versus Pressure Drop for Fibrous
       Bed	251
 16    Experimental Penetration Curve for Charged
       Droplet Scrubber 	  252

-------
                         PAPER 13

 1    Mean Differential Mass Loading Versus Aerodynamic
      Particle Size for All Venturi/Spray Tower System
      High Fly Ash Loading Runs	269

 2    Mean Differential Mass Loading Versus Aerodynamic
      Particle Size for the Venturi/Spray Tower System
      Low Fly Ash Loading Run	270

 3    Mass Percent Penetration'Versus Aerodynamic
      Particle Diameter for All Venturi/Spray Tower
      High Fly Ash Loading Runs	271

 4    Mass Percent Penetration Versus Aerodynamic
      Particle Diameter for Venturi Spray Tower
      Low Fly Ash Loading Run VFG-1B	272

 5    Particulate on Third Stage of Outlet Impactor
      During Venturi/Spray Tower Run VFG-1A,
      Magnification 2000X	273

 6    Particulate on Sixth Stage of Outlet Impactor
      During Venturi/Spray Tower Run VFG-1A,
      Magnification 5000X	273

 7    Mean Differential Mass Loading Versus Aerodynamic
      Particle Size for All TCA System High Fly Ash
      Loading Runs	274

 8    Mean Differential Mass Loading Versus Aerodynamic
      Particle Size for All TCA System Low Fly Ash
      Loading Runs	275
 9    Mass Percent Penetration Versus Aerodynamic
      Particle Size for All TCA High Fly Ash
      Loading Runs	276
10    Mass Percent Penetration Versus Aerodynamic
      Particle Size for All TCA Low Fly Ash
      Loading Runs	277


                         PAPER 14
 1    Waterloo Scrubber - Laboratory Installation	289

 2    Cross-Section Details of Waterloo Scrubber 	  290

 3    Dust Feeding Equipment 	  291

 4    Typical Particle Size Distributions (not normalized)
      in Inlet and Outlet Duct Samples on the Waterloo
      Scrubber	292

 5    Fine Particulate Removal Efficiency as Related
      to Fan  Speed	293


                             xi

-------
Figure                                                      Page

                      PAPER 14  (Continued)

  6    Water Droplet Size vs.  Air Pressure for Spraying
       Systems Air Atomizing Nozzle 	  294
  7    Number of Water Droplets Produced per Minute
       with the Spraying Systems Air Atomizing Nozzle .  .  .  295
  8    Histogram Showing Droplet Size Distribution from
       a Spraying Systems Air Atomizing Nozzle	296
  9    Fine Particulate Removal Efficiency as a Function
       of Dust Loading and Amount of Fine Water Spray
       Used	297
 10    Schematic Drawing of the Revised Consolidated
       Bathurst Installation	298
 11    Semrau Plot for the Recovery Boiler on a
       Kraft Pulp Mill	299
 12    Relationship Between Fan Speed and Capacity
       at Constant Residence Time	300


                          PAPER 15
  1    UW Electrostatic Scrubber	305
  2    General Layout of Electrostatic Scrubber Pilot
       Plant (Mark 2P Model)	306

  3    UW Electrostatic Scrubber Pilot Plant at Steel
       Plant	307
  4    Influence of Gas Residence Time on Particle
       Collection Efficiency	311

  5    Influence of Aerosol and Droplet Charging on
       Particle Collection Efficiency 	  312
  6    Influence of Water Droplet Charging Voltage on
       Particle Collection Efficiency 	  313

  7    Influence of Water to Gas Flow Rate Ratio on
       Particle Collection Efficiency 	  313

  8    Influence of SCA and L/G on Particle Collection
       Efficiencies 	  315

  9    Collection Efficiencies at an Electric Arc
       Steel Furnace	316
                             XII

-------
Figure                                                      Page
                          PAPER 16
  1    Hydro-Sonic Cleaners 	 330
  2    Energy Utilization Region	330
  3    Fan Coalescer	331
  4    Ejector Coalescer	331
  5    Shaped Droplets - Subsonic Nozzle	332
  6    Shaped Droplets - Supersonic Nozzle	332
  7    Droplet Growth 	 332
  8    Supersonic Ejector Drive	• .  .  .  . 333
  9    Coalescer - Energy vs Cleaning 	 333
 10    Fan Coalescer - Energy vs Outlet Grain Loading  .  .  . 334
 11    Paper Recovery Boiler	335
 12    Energy vs Particle Size	336
                            Xlll

-------
                       LIST OF TABLES
Table                                                       Page
                           PAPER I
  1   No.  1 Boiler House Scrubber Performance	15
  2   Houston Sinter Plant Steam-Hydro Performance 	   16
  3   Houston Sinter Plant Scrubber System Wastewater
      Characteristics	17
  4   Middletown Open Hearth Scrubber System
      Wastewater Characteristics 	   18
  5   Middletown Recycle Plant Scrubber System
      Wastewater Characteristics 	   19
  6   Middletown Recycle Plant Scrubber System
      Performance	20

                           PAPER 5
  1   F/C Scrubbing Studies by A.P.T. for EPA	45
  2   Process Streams for F/C Scrubber System at
      Metals Recovery Furnace	46
  3   F/C Scrubber Demonstration Plant Operating Modes ...   47
  4   Chemical Analysis of Particulates at Secondary
      Metals Recovery Furnace	48
  5   Chemical Analysis of Impinger Solution at
      Secondary Metals Recovery Furnace	49
  6   Cost Comparison for Premium Wire Recovery	50

                           PAPER 6
  1   Test Aerosol Characteristics 	  106
                          PAPER 10-A
  1   Mobile Particulate Collector Capabilities	191

                          PAPER 10-B
  1   Range of Conditions	197
  2   Sieve Tray Testing Conditions	198
                            xiv

-------
Table                                                       Page
                          PAPER 11
  1   Properties of Nikopol Electric Arc Flue Gas 	  214
  2   Results of Concentrations Measurements at Inlet
      and Outlet of Nikopol Scrubber	215
  3   Nikopol Scrubber Fractional Efficiency	216

                          PAPER 12
  1   Summary of Scrubbers Tested 	  235
  2   Design Equations for Various Scrubber Types 	  236

                          PAPER 13
  1   Flue Gas Characterization Program Results for the
      Venturi/Spray Tower System	266
  2   Flue Gas Characterization Program Results for the
      TCA System	267
  3   Selected Venturi/Spray Tower Outlet Mass Loading
      Solids Analysis and Resulting Calculated Reaction
      Products Emission 	  268

                          PAPER 15
  1   Source Test Parameters and Measurement Techniques .  .  308
  2   DOP Test Results	309
  3   Results of Tests at Coal-Fired Boiler 	  309
  4   Results of Tests at Electric Arc Steel Furnace.  . .  .  310

                          PAPER 16
  1   Scrubber Inlet Loading by Size Interval 	  327
  2   Scrubber Outlet Loading by Size Interval	328
  3   Inlet Particle Size Distribution	329
                             xv

-------
            INTRODUCTORY REMARKS
                     by
                Steven Reznek

   Office of Energy, Minerals and Industry
United Stated Environmental Protection Agency
           Washington, B.C. 20460

-------
                          INTRODUCTORY REMARKS
     Good morning.

     I would like to welcome you all to the second Fine Particle Scrubber
Symposium.  The first was three years ago in San Diego,  It is appropriate
that I give this talk today, since I represented EPA at the first (and
perhaps only) symposium on fabric filters.  So under the equal time
rule—let me now support scrubbers.  In keeping with the tradition of
conferences let me encourage everyone to exaggerate the attributes of
wet scrubbers, refrain from mentioning their shortcomings and ignore
competing control technologies.

     I want to address first the subject—why fine particulate?  As
you all know, nature designed your respiratory systems with a "wet
contact" scrubber.  Surface contact scrubbers are, of course, much more
efficient on large particles.  And your respiratory system is not effi-
cient at removing fine particles — those that are small enough to behave
as gases and follow gas flow, but are too large to be subject to molecular
diffusion.

     Because all particle control measures have similar limitations, the
particle sizes in controlled exhaust gases are likely to be at just that
diameter — about one or two microns — where they can most effectively
penetrate deep into the respiratory system.

     A fundamental question has long been whether control measures for
stationary sources are solving the less important problem of dust and
failing to affect the important problem of the particulate that actually
reaches the lungs.

     The problems of establishing the relationship between human health
and specific size and chemical fractions of particulate in the urban
atmosphere are enormously difficult and complex,,  Because of this com-
plexity it is very unlikely that air quality managers will soon feel that
particulate control technologies are fully available or completely
adequate.

     Therefore, we meet here today with shared interest both in increasing
the ultimate capability of control devices and in lowering the cost of
achieving a specified level of control.  So whether interest in improved
methods stems from the fear that we have not yet addressed the most
important health problem or from the conviction that while we can control
80 to 95% of the problem we would still like to reduce the remaining
emission, especially when particulates may be toxic or hazardous.  The
focus of research and development now is on particles a few microns in
diameter.

-------
     My  second question is why wet scrubbers?  Several forces are at work
which will result in a reexamination of wet scrubbers as the particulate
control  device of choice, other than the impetus toward controlling
smaller, more toxic particles.  Principally these relate to the required
control  of both gaseous and particulate- components in exhaust gases.  While
the primary focus of this conference is industrial applications, I can best
illustrate the point by referring to electric utility uses.  The desire
to control S02 and particulates has produced a technology - stack gas
desulfurization.  This technology, by proper design, can be optimized to
reduce the cost of S02 and particulate control,

     Another approach to meeting SOa standards is the use of low sulfur
coal.  The increased resistivity of the fly-ash, from low sulfur coal
reduces  the efficiency of electrostatic precipitators.  For high resis-
tivity particulates, scrubbers may present a viable alternative.  At
present, over 47,000 megawatts in electric generating capacity is com-
mitted to scrubbers -- principally for SOz control.
     The president's Energy Message strongly endorses increased coal use --
over 1 billion tons/year by 1990.  Because of the desire to assure the
transfer from an oil and gas based energy system will not entail a cost
in public health, EPA's energy program will support R£D on fine particu-
late control.  Part of that effort will be devoted to scrubbers.  While
the final program events await White House Staff approval to determine
exact amounts, fine particle scrubber research effects will be substantial.


EPA PROGRAM

     EPA initiated the Wet Scrubber Systems Study in 1970.  At that time
the state-of-the-art was largely empirical. Each application was considered
to be a special case.  Engineering design was based on primitive, try and
see approaches.  Very little scrubber performance information was avail-
able.  Often over-design increased the costs of operating units.

     In EPA's Wet Scrubber Systems Study all available information con-
cerning scrubber theory and practice was collected and evaluated.  En-
gineering design methods were evaluated and, where necessary, new methods
were developed to provide as sound a basis as possible for predicting
performance.  The result of this study was the 1972 publication of the
"Scrubber Handbook", now widely recognized as the most extensive engi-
neering text on the subject.

     Two recommendations for improved scrubber technology were developed
in the Wet Scrubber Systems Study:
1.   Study new phenomena which can enhance the ability to separate
particles from gas streams and;
2.   Study entrainment separators (or "mist eliminators") and develop
improved design methods and equipment.

-------
     Acting on the first, EPA began its effort for developing and demon-
strating Flux/Condensation scrubbers.  In a Flux/Condensation scrubber
water is condensed on the particles increasing their mass and thus making
separation easier.

     A series of condensation scrubbing studies have progressed from
theory, through experiment, and into industrial demonstration.   They have
produced a useful body of engineering knowledge.  It is now possible to
make reliable process designs and cost estimates for condensation scrub-
bers following inexpensive, small scale sampling studies.  For certain
emissions, condensation scrubbers can be economically superior to
conventional techniques.

     EPA has measured the fine particle scrubber performance for a number
of industrial sources.  These measurements are now the most detailed and
precise body of information available on collection efficiency as a
function of particle size.  All of EPA's scrubber performance studies
include a comparison of the measured efficiency with that predicted by
theory.  One very useful engineering design tool developed through the
EPA program is the relationship between particle cut diameter and required
scrubber pressure drop.  This relationship now allows reasonably accurate
estimates of the power required to achieve specified collection effi-
ciencies.

     The second recommendation concerned re-entrainment.  The problem of
entrainment has placed serious constraints in the overall efficiency of
scrubber operation, particularly for lime/limestone slurry systems.  A
portion of EPA's efforts have been aimed at this problem.  The results
have been knowledge of the efficiency, capacity and pressure drops of
entrainment separators, and of the size distribution of the entrained
mist.  Here again, the EPA program has expanded the rational basis for
the engineering design of better, more reliable and more economical
equipment.

     Some of the research problems remain for wet scrubbers.  Principal
among these are re-entrainment and reduced overall removal efficiency,
the energy penalties of scrubbers including stack gas reheating and the
overall environmental impact of scrubber blow down, its handling and
disposal.  However, scrubbers are perhaps the only technology applicable
to corrosive or dangerous stack gases.  Another obvious advantage of
scrubbers is their ability to be used for both particulate and gaseous
po1lutant removal.

     This symposium is intended to stimulate and generate new ideas for
fine particulate control by bringing together experts in the development
and use of wet scrubbers.

     I want to wish you well in your efforts both here and out there in
the French Quarter0

-------
        WET SCRUBBING EXPERIENCES FOR STEEL MILL APPLICATIONS
                          B. A. Steiner
                          R. J. Thompson
                      Armco Steel Corporation
                        Middletown, Ohio
                           ABSTRACT
     This paper will  discuss wet scrubbing experiences  at  a
variety of steel mill applications including sinter plant,
blast furnace, open  hearth,  and industrial boiler  installa-
tions.  A number of  case  studies will be examined.  For each,
the process, emission characteristics, and wet scrubber sys-
tem design will be described.   Actual performance  will  be
compared with design  values, and particular emphasis will  be
placed on water chemistry control and related problems.

-------
           WET SCRUBBING EXPERIENCES FOR STEEL MILL APPLICATIONS

     Discharges from air pollution sources in the steel industry are often
characterized by high concentrations of finely divided particles, high gas
temperatures, moisture, and the presence of certain troublesome gaseous
pollutants.  Some process discharges also contain sub-micron condensed hydro-
carbons, sublimates, or carbon particles which make gas cleaning more diffi-
cult.  When these characteristics are complicated further by wide variations
and fluctuations associated with the cyclic nature of many steelmaking
processes, it is easy to comprehend why wet scrubbers, which are well-suited
to cope with many of these design requirements, have been so widely used in
the steel industry.  Virtually every major type of air pollution source in
the industry has been controlled by a wet scrubber application at one plant
or another.

     Among all major types of air pollution control equipment, the wet
scrubber represents the smallest investment as a proportion of the total
pollution control system cost.  For example, the installed cost of the wet
scrubber and mist eliminator portion of a scrubbing system may represent only
5-10$ of the total system cost.  In contrast, a fabric filter or electro-
static precipitator may represent 25-30$ of the total system cost.  Although
auxiliary equipment is important for all pollution control applications, the
proportionately greater emphasis on auxiliary systems for wet scrubbing
systems substantially increases their significance and the dependence on
proper design and operation.  The interface with water pollution control
requirements inherent with wet scrubbing systems also results in a greater
emphasis on the waste treatment aspects of a wet system.  Therefore, it comes
as no surprise that many of the unfavorable experiences with wet scrubbing
systems are related to the auxiliary systems and in particular the water and
waste treatment aspects of the system.

     Armco Steel Corporation has installed over l+O major scrubber systems in
the past 10 years, including blast furnace, basic oxygen furnace, open hearth,
electric furnace, scarfer, coal dryer, sinter plant, coke plant, and boiler
applications.  Several of these systems have been plagued with problems of
varying magnitude which will be discussed as case histories below.

Middletown Coal-Fired Boilers

     The No. 15 and No. 16 boilers at the Middletown (Ohio) No. 1 Boiler
House are multi-retort, underfeed stoker, coal-fired units, each with a capa-
city of 100,000 pounds of steam per hour.  Although not an application unique
to the steel industry, this source typifies many of Armco's scrubber install-
ations.  The coal burned is typically 0.7-0.9$ sulfur and 8-10$ ash.  This
boiler house is normally operated in the winter months only.

     A scrubber was selected for this application to provide removal of sul-
fur dioxide as well as particulate matter to allow burning of high sulfur
coals if necessary.  A single scrubber was installed in 1975 to serve both
boilers.  Gases are extracted from the existing stacks and passed successively
through a fan, scrubber, mist eliminator, and stack (Figure 1).

     The scrubber consists of a variable throat, orifice venturi scrubber
with a pressure drop-controlled adjustable plug.  The 17' diameter mist
eliminator is equipped with three gas scrubbing trays, and "S" type chevron

-------
separator, and an 8'-10" diameter stub stack.  The system is designed for
11+0,000 acfm @ I|.25OF and 15" w.c. with an outlet loading of 0.02 gr/dscf.

     With the present design, fresh water is supplied from a clarified river
water system at a rate of 1600 gpm to both the top tray of the demister and
to the venturi throat.  Water from the mist eliminator flows into a 10,000
gallon recycle tank where lime slurry is added to maintain an approximate pH
of 6.  Water from the tank is recirculated to the bottom two trays of the
demister.  Waste water from the tank is pumped to an abandoned hot strip mill
flume and scale pit for settling of solids.

     The largest problem affecting scrubber performance has been control of
excess air on the boilers.  Without control, excess air in the range of 100-
250$ was experienced, resulting in 1-^-3 times the scrubber design flow rate.
At this flow the scrubber and fan were 'unable to maintain the 15" w.c.
pressure drop while providing proper boiler stack draft.  These high excess
air rates also resulted in the presence of unburned carbon which could not be
wetted and compounded the poor performance.  By controlling excess air in the
range of 60-80$, gas volume was kept within reasonable limits, permitting
the plug to adjust to the proper location, thereby providing acceptable
scrubber efficiency (Table l); however, the design outlet grain loading has
yet to be obtained.

     The originally installed mist eliminator chevrons were constructed of
Norel plastic rated at 225 F continuous duty, which was well above the temper-
ature at that location in the system.  However, overlooked was the fact that
Norel is creep sensitive at 20 F above ambient and the chevrons failed within
a few months.  New chevrons of the same material are also being damaged.
Either stainless steel or fiberglass reinforced plastic (FRP) is a preferable
material for this application.

     The original water system was designed to apply fresh water to the de-
mister trays, and to use recirculated water with a suspended solids concentra-
tion of 1|0-50,000 mg/1 at the venturi.  All piping under this service was mild
steel and was severely abraded, causing numerous leaks.  The piping was
replaced with the same material, but was modified so that the water to the
venturi is now once-through, and the abrasion has ceased.

     A similar leakage problem was caused by abrasion of the 12" steel pipe
carrying dirty water from the scrubber trays to the recycle tank.  Partial
flow in this pipe, which caused a wet-dry line, was also a contributing
factor.  This line was replaced with a rubber-lined flanged elbow and rubber
hose, which have held up very well to date.

     Abrasion also caused leaks in the pumps which deliver sludge from the
recycle tank to the waste treatment system.  Lowering pump speed from 1780 to
1180 rpm helped greatly to reduce abrasion.  In addition, rubber lined,
standard flanged "T" sections were installed on the suction side of the sludge
pumps and recirculation pumps and have withstood abrasion better than any
other material tried.

     Generally speaking, it is believed that all flyash slurry lines should be
rubber-lined and assembled in lengths of ten feet or less for ease of handling.

                                      7

-------
     One problem which was anticipated but has not occurred was lime buildup
in the demister trays during low gas flow rates.  A pH of l;-6 was thought
necessary to minimize this problem.  However, at sustained pH levels of 6-6.5
there has been no significant lime accumulation on the trays.

     The lime slurry system has also presented problems.  Clogging of the
lime slurry line is attributed to its excessive length.  Control of the lime
addition rate is also a problem, but can be resolved with the installation of
a proportional feed lime slaker.

Houston Sinter Plant

     The Armco Houston Works Sinter Plant produces a sized agglomerate for
blast furnace feed from finely divided iron bearing material.  Coke breeze is
used as a fuel and limestone as a flux.  These materials are mixed together
with a blend of iron ore fines, blast furnace flue dust, blast furnace filter
cake, and mill scale, and are placed in a 10-12 inch layer on a traveling
grate.  The surface of the bed is ignited with natural gas and progressively
burns downward as combustion air is drawn through the bed from top to bottom
through a series of windboxes and an induced draft fan.  The fused material
is sinter which is subsequently crushed, screened, cooled, and screened again
prior to use in the blast furnace.  The plant produces 1500 tons per day of
sinter at a base/acid ratio of 1.0 to 1.5.

     The plant was originally installed with cyclonic separators for removal
of particulate matter ahead of the fan.  The resultant discharge was about
0.3 gr/scf dry, which was in excess of Texas process weight limitations, as
well as opacity restrictions of 30$ .  About l+^fo of the particulate matter was
determined to be chloroform-extractable hydrocarbon, which was driven from
oily feed material and condensed.  These hydrocarbons were a primary contribu-
ting factor to stack opacity and represented a major consideration in
selection of a control system.  Also present in the stack discharge were
sulfur dioxide, chlorides, and fluorides.

     Five different control systems were investigated, including dry and wet
electrostatic precipitators, fabric filters, wet scrubbers, and Lone Star
Steel's Steam-Hydro scrubber.  Both high energy venturi scrubbers and the
Steam-Hydro scrubbers were pilot tested.  Both were found to satisfactorily
remove particulate and, to a lesser extent, hydrocarbons; however, the Steam-
Hydro was considerably more efficient on hydrocarbons.  Because of the con-
cern for opacity, the high hydrocarbon efficiency gave an edge to the Steam-
Hydro system, which also had a slight economic advantage, but only due to the
availability of steam generating capacity and low cost fuel to generate the
steam.

     The Steam-Hydro system was installed in 1975 as six parallel units, five
operating and one spare.  Each unit consists of (l) a combination steam-water
nozzle where the steam atomizes the water droplets, (2) a mixing tube where
the particles and droplets contact, and (3) twin cyclone separators where the
particle-laden droplets are removed and the clean gas is discharged through
stub stacks.  The original induced draft fan supplies the dirty gas to the
units, and the motive force of the steam forces the gas through the gas clean-
ing system.  Results of performance tests on the system are shown in Table 2. .
Opacity problems still exist periodically due to high oil content of sinter
feed materials.
                                      8

-------
     Waste water from the Steam-Hydro cyclones is neutralized and clarified
in two parallel systems, each capable of treating the entire flow.  Each
system consists of a lime neutralization unit, flocculator, clarifier, and
sludge pump (Figure 2).  Hydrated lime is fed from a storage bin and volu-
metric feeder into a mechanical mix tank.  The neutralized waste flows to a
Lamella separator equipped with an integral flash mixer and flocculator.
Polymer additions are capable of being made but have not been necessary.  The
Lamella underflow is pumped to the blast furnace gas cleaning water system
for ultimate disposal.  The Lamella effluent is recirculated to the Steam-
Hydro units at a rate of 370 gpm.

     The principal problem with the system since start-up has been weld
failures in the 30U stainless steel components.  These failures were
corrected in the field and have continued to occur, but at a much slower
rate.  Fabrication practices and corrosion may be contributing factors.

     Also, steam expansion at the steam-water nozzles results in considerable
noise.  Measurements near the units indicate levels of 110 dB on the A scale.
However, the noise is very localized, and because no operator is required in
the immediate area, no corrective action has been necessary to date.

     Several problems have been encountered in the water system, primarily in
the lime handling and sludge handling areas.  Self-cleaning pH probes were
originally installed in the neutralization tanks for automatic control of lime
feed.  At this location the probes were rapidly coated with calcium sulfate
scale and malfunctioned.  This problem was solved by relocating the probes to
the effluent end of the Lamella separators, but frequent cleaning is still
required.

     The lime storage and feeding equipment is located above the neutraliza-
tion tank.  Dry hydrated lime falls from the feeder through a vertical 2-inch
pipe directly into the tank.  Because of the high humidity in the area, the
lime feed pipe plugged frequently.  Although plugging has not been eliminated
completely, it has been reduced to an acceptable frequency by (l) adjusting
the pH control system so that lime feeding is nearly continuous, and (2) blow-
ing dry air through the lime feed pipe to minimize the amount of moisture
entering the pipe.

     Air operated diaphragm pumps were initially installed to transfer
Lamella underflow slurry to the blast furnace water system.  Numerous dia-
phragm failures were encountered with these pumps.  After a year of unsuccess-
ful attempts to find a reliable diaphragm material, these pumps were replaced
with centrifugal pumps.  No problems have yet been experienced with the new
pumps.

     Shortly after the system was started up, corrosion coupons were in-
stalled.  The first test indicated a corrosion rate of 68 mils per year (mpy)
on carbon steel and no corrosion on 30U stainless steel.  Although the entire
gas cleaning system is constructed of 30^ stainless steel, the entire water
system is carbon steel construction.  The high corrosion rate was attributed
to chlorides and sulfates in the waste water and poor pH control due to the
problems previously discussed.  After a more stable pH control operation was
achieved, a second test indicated corrosion rates of 10 to 20 mpy on carbon

-------
steel.  Further refinements in the pH control system have been made, and a
third corrosion rate test is now being conducted.  If the results of this
test continue to indicate excessive corrosion, chemical corrosion inhibitors
will be investigated.

     It should be noted that neither the slurry conveyed to the blast furnace
water system nor the Lamella effluent is acceptable for direct discharge to
a stream.  The treatment facilities described are designed for making the
water amenable to recirculation only.  Table  3   compares the characteristics
of the recirculated water, a portion of which might ordinarily be blown down
to a stream, with tentative EPA Effluent Guidelines.  The differences repre-
sent the required removal of pollutants before direct discharge can occur.
These additional facilities are likely to experience similar, perhaps more
complex, difficulties in operation and maintenance.

Middletown No. 2 Open Hearth

     The Middletown No. 2 Open Hearth Shop consists of six 300-ton furnaces,
each with a maximum firing rate of 170 MMBTU/Hr.  Oxygen is normally used at
rates of 60,000 scf/hr with a maximum of 100,000 scf/hr.  Checker blowing is
conducted on a I|.-hour cycle.  The furnaces were designed to be fueled by 75%
natural gas and 25% "tar.  However, since 197^4- with severe gas curtailments,
they have been fired primarily with tar and fuel oil.  These furnaces are now
normally on a 100% scrap charge which results in average heat duration of
about 12 hours.  Depending upon availability of molten pig iron, the furnaces
can also be charged with hot metal which significantly reduces heat time.

     Scrubbers were installed at this facility in 1970.  Scrubbers were
selected because of greater expected reliability and flexibility over electro-
static precipitators.  There was also a decided cost edge because precipita-
tors would have required the added expense of installing waste heat boilers.

     The scrubbers are orifice type, adjustable throat, venturi scrubbers of
the flooded disc design.  Each furnace has an individual scrubber system
consisting of venturi scrubber, flooded elbow, cyclonic separator, fan, and
a sound-attenuated stack (Figure 3)-  The open hearth stack gases have a
maximum flow rate of 225,000 acfm @ 1350 F with an average dust concentration
of 5 gr/scfd.  During oxygen lancing, the dust concentration is about
6 gr/scfd, and during checker blowing concentrations as high as 25 gr/scfd
may be present.  Depending on the fuel being used, sulfur dioxide concentra-
tions may range from negligible to 650 ppm.  The guaranteed outlet dust
concentration is 0.03 gr/scfd.  A series of 19 performance tests in 1971 at
a scrubber pressure drop of 14.7" w.c. showed an average outlet particulate
concentration of 0.0127 gr/scfd.

     Scrubber water is supplied from and returned to a single water treatment
plant.  Within each scrubber system, water is recirculated to the flooded
disc from a pump tank at a rate of 1000 gpm.  An additional 500 gpm of makeup
water is supplied from the treatment plant to the wetted inlet of each
scrubber.  With approximately 200 gpm of evaporation, the remaining 300 gpm
is returned to the treatment plant for pH adjustment and clarification.  At
the treatment plant, a blowdown of about 100 gpm is made to control dissolved
solids, and fresh makeup is added.

                                      10

-------
     The system was originally intended to be operated without pH control,
and all components were constructed either of 316L stainless steel or of
fiberglass reinforced plastic (FEP).  Because of the presence of sulfur,
chlorides, and fluorides and the resulting formation of acidic solutions, the
system initially operated with recirculated water with a pH of 2 to 3-
Several problems were encountered.  Mist carryover from the stack was acidic
and fallout in the area was of concern.  Stress corrosion cracking of some
stainless components and total corrosion of some mistakenly installed non-
stainless hardware was occurring.  Also, pH adjustment of the final blowdown
was causing precipitation of large quantities of solids which were soluble
at the lower operating pH levels.  These solids were creating severe plugging
problems in equipment.

     To circumvent these problems, a lime neutralization system was installed
to maintain a scrubber water pH of 8 to 8.5>.  Experience under this mode of
operation also was unfavorable due to widespread scaling of calcium sulfate,
especially when higher sulfur fuels were used for firing the furnace.  In-
creased iron precipitation was also occurring in the system.  Due to the
presence of chlorides and fluorides, acute acidic attack was occurring in
some areas while the severe scaling was occurring in other areas.  More
recently, caustic has replaced lime for neutralization, a pH of 5-5 to 6 is
maintained, and the scaling problem is mitigated to some degree.

     Corrosion was worst in the 316L stainless steel clarifiers.  The clari-
fier sludge rakes were patch-welded and the interiors of the clarifiers were
sand blasted to remove scales, then coated with a coal tar epoxy.

     All water piping in the system is PEP.  Proper design allowances were
not made for the high thermal expansion rates of PRP piping, and it was nec-
essary to install additional expansion joints.  Plugging of the piping was
also a problem and was attributed to the failure to use large radius elbows.

     At the scrubbers, mist eliminator water was intended to be discharged to
a screw classifier to remove large particles before recirculating or return-
ing the water to the treatment plant.  These classifiers were largely
ineffective because of the small particle size of the solids and were removed.

     The positioning mechanism for the scrubber malfunctioned due to the
accumulation of sludge in the packing box as a result of leakage of dirty
water around the top seal of the box, causing the shaft to bind.  To allevi-
ate this, the bottom seal was removed to allow the sludge to drain from the
packing box, and wear plates were added to the shaft.

     Numerous cracks appeared in the welds and blades of all the fans
shortly after start-up.  These problems were extremely critical at the time
and after many tests, were attributed to a marginal stress application for
316L stainless steel, which had not previously been used for this size and
type of fan.  The fans were redesigned and modified in the field and have
since experienced no problems of that nature.

     One other water problem has been the occasional occurrence of a foam in
the treatment plant.  Although still largely unexplained, it has been
suggested that, because of a high concentration of zinc oxide (63$) i-n

                                     11

-------
foam, the zinc is causing a change in surface tension of the water.  The
sporadic nature of the problem may be due to the fluctuations in zinc scrap
additions in the open hearth furnace.

     Table  4 shows the present treatment plant discharge in comparison to
tentative EPA Effluent Guidelines.  Modifications to the plant are now under-
way to achieve the BPCTCA limits.

Middletown Recycle Plant

     In 197^» a new plant was built in Middletown for the recycling of iron-
bearing waste materials.  The facility is referred to as the Recycle Plant
because it is intended to convert exclusively in-plant iron-bearing waste
fines into blast furnace feed material.  The process is very similar to the
Houston operation but is designed to produce 261;0 tons/day of sinter at a
base/acid ratio of 3-0 to 3«5«

     Based upon wet scrubber pilot plant testing at two existing sinter plants
and concern for hydrocarbon emissions, a high energy, variable throat, orifice
venturi scrubber with an adjustable plug was selected.  The gas cleaning
system consists of, in sequence, dry primary and secondary cyclones, two in-
duced draft fans in series, a wet scrubber, and a cyclonic mist eliminator
with a stub stack (Figure U).  The system is designed for 30l±,500 acfm @ 300°F
with a venturi pressure drop of l^" w.c. with an outlet loading of 0.02
gr/scfd.

     Vater is supplied to the scrubber throat and inlet at 1800 gpm recircu-
lated from the mist eliminator tank.  Make-up water is supplied to the mist
eliminator at the rate of lj.00 gpm from other sinter plant noncontact cooling
water sources.  From the bottom of the mist eliminator tank, a blowdown of
325 gpm is pumped to blast furnace sludge settling ponds and becomes part of
the blast furnace scrubber water system blowdown.  A hydrated lime neutraliza-
tion system is provided for pH control of the recirculated water, but the
system has been operated in the acid mode without problems, and no neutraliz-
ing has been necessary.  The wastewater characteristics of the blowdown are
shown in Table 5.

     The major problem with this installation has been the design of the
scrubber itself.  The initial design exhibited a pressure drop of 80" w.c.
Following some field alterations, including leveling of the weir which
supplies water to the inlet, proper alignment of the bull nozzle which
supplies water to the plug, and widening of throat opening, the scrubber drop
still could not be lowered below 55" w.c.  Even at this pressure drop, the
outlet grain loading was reduced only to .07-.09 gr/scfd.  As a result, the
scrubber manufacturer supplied a second scrubber design with a modified throat
profile.  Although the second design resulted in better performance, an
operating pressure drop of 68" w.c. was required.  Mechanical difficulties,
inability to properly control pressure drop, and an inadequate throat opening
all plagued this second design.  A third design with yet a larger opening and
a modified throat profile was furnished by the manufacturer in early 1976,
and performance and pressure drop are now in accordance with design criteria
(Table  6). These problems were all believed to be attributed to the scrubber
manufacturer's basic unfamiliarity with the capabilities of its own scrubber
designs.
                                      12

-------
     Some performance difficulties were experienced even with the third
scrubber design with respect to removal of condensed hydrocarbons.  Initial
tests revealed grain loadings in excess of .07 gr/scfd and excessive stack
opacities.  Pilot testing had previously shown scrubbers to be capable of
some removal of hydrocarbons but with certain inlet loading limitations.
Only after selective elimination of certain excessively oily feed materials
were the results in Table VI able to be achieved.

     A number of mechanical problems were also experienced but all were
attributed to marginal mechanical design rather than corrosion or abrasion
problems characteristic of the other case histories discussed previously.
These problems included failure of shaft bushings, cracked expansion joints,
collapsed water hoses, failure of mist eliminator core buster supports, and
peeling of mist eliminator protective lining.  Bushings were replaced, shaft
diameter was increased, hoses were replaced with a stronger material, ex-
pansion joint designs were modified, and the mist eliminator lining was
patched with a better material.

     The water system has been generally free of problems for the following
reasons:  (l) proper material selection based on other system experiences,
 2) lack of need for chemical additions, (3) lesser quantities of acid gases,
 k) greater blowdown percentage, (5) fan placement upstream of the scrubber,
and (6) greater mechanical collector efficiency ahead of the scrubber.

Conclusions

     These four case histories illustrate some of the types of problems
experienced by Armco Steel Corporation with wet scrubbing installations.
These experiences have aided Armco with subsequent system designs, and it is
hoped others will also benefit from the descriptions of these problems.  Al-
though most problems have been resolved or corrected, one of the basic
reasons for these difficulties persists.

     USEPA Effluent Guidelines place emphasis on minimization of waste water
discharge flows and attainment of best available technology.  To achieve
these standards, maximum recirculation is required.  As recirculation rates
are increased, many of the types of problems discussed come into play.
Chemical treatment and solids removal processes and equipment become more
sophisticated.  The accumulation of dissolved solids and acidic solutions
and the attendant fluctuations in pH lead to scaling and/or corrosion prob-
lems.  Suspended solids buildups cause abrasion and erosion.

     More importantly, however, only the very basic of the water treatment
needs have been addressed to date.  As efforts to remove such pollutants as
fluorides, cyanides, phenol, ammonia, and heavy metals are required by BATEA
water standards, the air pollution control technology of scrubbers is likely
to be overshadowed by water pollution control technology for the scrubber
water.  Although there are several reasons that scrubbers are losing favor
for steel mill applications, energy considerations and competition with
other more efficient control devices among the most important, the concern
for water and waste treatment standards also looms as a significant factor.
When New Source Performance Standards are proposed with scrubbers considered
to be best available control technology, the water treatment aspects cannot

                                      13

-------
"be casually dismissed as an engineering detail.  Development of waste water
treatment technology must simultaneously accompany scrubber technology if
scrubbers are to remain as an influence in the air pollution control field.
                                      14

-------
                        TABLE I
          No. 1 Boiler House Scrubber Performance

No. Tests                                   3
Ave. Steam Output                     191,571  Ib/hr
Ave. Heat Input                            252  MMBTU/hr
Ave. Excess Air                             71  % (V)
Ave. Moisture Content                      5.3  % (V)
Ave. Temperature                           71  °F
Ave. Flow Rate                         83,233  SCFM dry
Ave. Particulate Loading                 0.0465  gr/SCF dry
                                       0.130  Ib/MMBTU
Allowable Particulate Loading             0.152  Ib/MMBTU
                           15

-------
                        TABLE II

       Houston Sinter Plant Steam-Hydro Performance
No. Tests
Sinter Production
Average Inlet Conditions
  Volume
  Temperature
  Moisture
  Particulate*
Average Outlet Conditions
  Volume
  Temperature
  Moisture
  Particulate"

Allowable Discharge*
Average Inlet SO2 (3 Tests)
Average Outlet SO2 (3 Tests)
   58.8   Tons/Hour
134,350   SCFM
    167   op
    4.7   %(V)
 0.5011   gr/scf
134,250
147
20.8
0.0301
38.5
90.6
66.8
10.0
SCFM
% (V)
gr/scf
Ib/hr
Ib/hr
ppm
ppm
*Total Particulate, including filter and impinger catch, State of Texas
 Compliance Sampling Manual
                             16

-------
                           TABLE III
              Houston Sinter Plant Scrubber System
                    Wastewater Characteristics
                                                    USEPA
                                 Recirculated      Effluent Guidelines
Parameter                           Water       BPCTCA     BATEA
Discharge, gal/ton                    44*        50         50
pH                                 8.5        6-9        6-9
Suspended Solids, mg/1              350         50         25
Oil & Grease, mg/1                   —           10         10
Fluoride, mg/1                       —           —          20
Sulfide, mg/1                        —           —         0.3
Dissolved Solids, mg/1             4300         —         —
Chloride, mg/1                       720         —         —
Sulfate, mg/1                      1900         —         —
Ammonia (N), mg/1                  144        125**       10**
Cyanide (CN), mg/1                <0.1         15**      0.25**
Phenol, mg/1                        3.0          4**      0.5
 * Assumed discharge flow, equivalent to the present slurry flow.
**Blast Furnace guidelines, all others are Sinter Plant.
**
                              17

-------
                                   TABLE IV
                    Middletown Open Hearth Scrubber System
                           Wastewater Characteristics

                                                                   USEPA
                              Existing     Water Quality-Based      Effluent Guidelines
Parameter                     Slowdown       NPDES Limit       BPCTCA     BATEA
Flow, gal/ton                    44             —             50        50
pH                              6.5            6-9             6-9        6-9
Suspended Solids, mg/1          95             34             50        25
Fluoride, mg/1                   400             22             —         20
Nitrate, mg/1                    28             —             —45
Zinc (Total), mg/1              1940             10             —          5
Lead (Total), mg/1               12              1             —         —
Iron (Total), mg/1                44             —             —         —
Iron (Diss.), mg/1                —                2             —         —
Sulfate, mg/1                  3200             —             —         —
                                       18

-------
                           TABLE V
           Middletown Recycle Plant Scrubber System
                  Wastewater Characteristics
                             Recycle Plant
Parameter                      Slowdown
Flow, gal/ton                    177
pH                              4.5
Suspended Solids, mg/1          440
Oil & Grease, mg/1               135
Fluoride, mg/1                    55
Sulfide, mg/1                      2
Sulfate, mg/1                    520
Ammonia (N), mg/1               14
Cyanide (CN), mg/1              0.13
Phenol, mg/1                     2.2
Zinc, mg/1                        0.3
      USEPA
 Effluent Guidelines
BPCTCA
  50
 6-9
  50
  10
 125*
  15*
   4*
BATEA
  50
 6-9
  25
  10
  20
 0.3

  10*
0.25*
 0.5*
   5**
 *Blast Furnace Effluent Guidelines
**Steelmaking Effluent Guidelines
                              19

-------
                     TABLE VI
              Middletown Recycle Plant
            Scrubber System Performance
No. Tests
Ave. Sinter Production
Ave. Process Weight Rate
Ave. Gas Flow
Ave. Temperature
Ave. Moisture
Ave. Particulate Emission*
Allowable Emission
      3
  101.6
  159.3
200,000
    130
   12.5
 0.0173
   0.29
   0.19
  29.58
  50.00
T/Hr
T/Hr
SCFMD
°F
% (V)
gr/SCFD
Ib/T sinter
Ib/T feed
Ib/Hr
Ib/Hr
*USEPA Method 5, front half only
                         20

-------
O
                                21

-------
•a   3
 •  to
;D  ti

-------
23

-------
24

-------
                 WET SCRUBBING EXPERIENCE
                   WITH FINE BORAX DUST
                       E. Dean Lemon
                U.S. Borax and Chemical Co,
                     Boron, California
                         ABSTRACT

     Borax dust emissions from borax fusion furnaces and
calciners can be controlled by a moderate pressure drop
scrubbing system designed to fit the particle characteris-
tics.  A case history is given for the upgrading of an
inadequate scrubber system to a successful one by modifi-
cation of existing equipment.  Power requirement for the
upgraded scrubber is 400 H.P., as compared to 700 H.P. for
the alternative high energy scrubber which had been suc-
cessfully used on other furnace lines.  Capital cost for
the modified system was much less than that of a new
replacement.
                            25

-------
                 WET SCRUBBING EXPERIENCE
                   WITH FINE BORAX DUST
     Many of us have been confronted in recent years with the
problem of what to do in the face of increasingly stringent
emission requirements with emission control equipment giving
marginal or inadequate performance for the revised standards.
In many cases, we have been required to go to new equipment,
or new technology and have had to rip out equipment still
capable of performing but not at the level required.  This
paper is primarily a description of a solution to this prob-
lem by modification of existing equipment and it indicates
that, given the time and with application of ingenuity and
effort, it may be possible to devise a better and cheaper
approach through modification or retrofit.
     In 1958, our Company completely changed its mode of
operation from underground to open pit mining and, in con-
nection with that change, revised its primary processing
from dry concentration to a wet solution and recrystalliza-
tion process.  A secondary process for manufacture of Anhy-
drous Borax was incorporated in the change with a method and
production capacity different and larger than anything we
had attempted previously.  This process, which is referred
to as fusing, generated a serious emission problem.
     In the late 1960's, when it became necessary to add to
original capacity, it also became necessary to install addi-
tional control equipment.  For  various  reasons, wet scrubbers
were selected.  After a look at the requirements of the then
young Los Angeles Air Pollution Control District, which covered
the next adjacent county, particulate emission standards of
0.3 gr/SCF were set as an objective and impingement type,
comparatively low energy scrubbers were installed to attain
it.  Our presumption that future regulations for a remote,
unpopulated desert area would never be more demanding than

                            26

-------
those for the urban area proved incorrect.  In the early 1970's
the plant was faced with 0.2 gr/SCF standards for existing
equipment and 0.1 gr/SCF for new equipment and, most critical,
a process weight rule  (the so-called 40# rule) that set par-
ticulate emission standards for the current operating condi-
tions at a maximum of  18 pounds per hour.  The previously
installed scrubbers met the 0.2 gr/SCF requirement but pro-
duced emissions at 34  to 62 pounds per hour.
     Further capacity  increases in the early 1970's success-
fully incorporated high energy (high pressure drop) scrubbers
so it was known that this technology could be used.
     The borates primarily used in commerce are the sodium
salts of boric oxide with varying degrees of hydration span-
ning the range from zero (or anhydrous) to 10 mol (or 47% water).
Sodium borates occur naturally in two forms, both of which
are found in the open  pit mine at Boron, Calif.  They are
Tincal, the 10 mol hydrate, and Kernite, a 4 mol hydrate.
Processing of these ores consists of dissolving the borate
portions, separating the insoluble gangue and recrystallizing
the borates.  Application of the phase rule and temperature
controls causes crystals to form as sodium borate decahydrate,
NaaBitC^ • lOHaO, which is 471 water, and as sodium borate pen-
tahydrate, NaaB^O?•5H20, 31% water.  A third sodium borate,
anhydrous borax, is produced by fusing one of the hydrated
forms, that is, heating it above the melting point of sodium
borate, which is 742°C (1,360°F).
     The hydrated borate crystals are subject to intumescence,
that is, swelling on rapid or severe heating.   This is caused
by the rise in vapor pressure inside the crystal to a point
where the crystal is ruptured and the water escapes.   The
ruptured crystal takes on the appearance of popcorn and is
very light and easily  friable,  breaking down readily into
fine, light particles.  This material can have an apparent
bulk density as low as one pound per cubic foot and will run
20 to 30% sub-micron particle size if it has been subjected
                            27

-------
to much handling.  These particles are white and even the
finest particles diffuse and scatter light so that visibility
of this material as a particulate emission is very high.
     The process for manufacture of Anhydrous Borax has been
the source of our worst dust problem.  Beginning with 10 mol
borax, the material is put through rotary kilns where it is
heated by the off gasses from the fusing furnace and calcined
to about 3 or 4 mols of water content.  It is then introduced
to the fusing furnace where it is melted under temperatures
of 1,800° to 2,000°F.
     There are five furnace lines using the process.  Furnace
capacities vary from 80 tons per day to 100 tons per day.  Air
quantities required average 45,000 ACFM.  Flue gas introduced
to the kiln is discharged from the furnace at 1,800°F, is
diluted with outside air in a mixing chamber, and the flow
rate through the kiln is approximately 10,000 ACFM at 800°F.
The gas stream arrives at the dust control equipment at a
temperature of 350°F to 500°F.  Heating in the kiln is rapid
and the vapor pressure within the pieces of borax becomes
extreme.  Intumescence is, therefore, severe and the potential
for creation of fine particulates is extremely high.  Temper-
atures in the fusion furnace are so high that there is vapor-
ization and condensation leading to fume formation.

Control Efforts
     In connection with a fusing furnace much smaller but
otherwise similar to those presently used, we successfully
installed a fabric type dust collector, using glass fiber
material for the bags because of the high temperatures. After
some difficulty in obtaining satisfactory operation due pri-
marily to maintenance problems, this collector has been made
to work satisfactorily from an emission control standpoint.
However, the size of collector required for the larger fur-
naces has obviated this approach for the new installations.
                            28

-------
     In 1958 the first four of the large furnaces were built
and operated using only cyclone type dust collectors.  This
proved to be highly unsatisfactory and within two years ef-
forts were begun to install additional control equipment.
After some test work, it was decided that wet scrubbers would
                                /
be the most practical for this installation due primarily to
the proximity of plentiful supply of process water and because
area limitations would not allow construction of amply-sized
baghouses.  As stated earlier, a particulate standard of 0.3
grains was selected as the objective.  A fifth furnace was
built in the early 1960s and the low energy scrubbers were
installed on all five.  These scrubbers were an impingement
type operated at a 14" w.c. pressure drop and handled an
average 45,000 ACFM with 200 H.P. fans.  These fans had to
be run wet to minimize buildup of borax deposits on the fan
blades.  By contrast, three high energy scrubbers installed
later with larger furnaces operated at 35" and 40" w.c. pres-
sure drop, handled 70,000 ACFM to 100,000 ACFM and used be-
tween 1,000 connected horse power and 1,500 connected horse
power.
     After an Air Pollution Control District was established
for the local area a process weight rule was adopted and the
five oldest furnace lines were in violation so we investigated
possible solutions to the problem.  An obvious approach was
to install the same type scrubber with 40" w.c.  pressure drop
as had been successfully used on the latest installations.
At that time, the first rumblings of energy shortage were
being heard and it was decided that something less energy-
demanding than the 1,000 to 1,500 H.P. fan motors should be
sought.
     A first attempt was the installation of knitted-mesh
mist eliminators.  Experimental work had indicated that some
of the escaping particulate material was dissolved in mists
escaping from the stacks.  This proved to be effective at
times but was not reliable and had to be rejected.
                            29

-------
     Several alternatives were then considered, among which
was a further modification of the impact type scrubbers.
A.P.T. Inc. was retained to study the problem and to deter-
mine whether or not scrubber performance could be improved
to give satisfactory results with lower energy requirements
than the high pressure drop scrubbers and less capital in-
vestment than other total replacement methods.
     The process gas streams were sampled at various loca-
tions to determine particle size, particle  concentration, gas
flow rates, and  other parameters.  Particle mass growth due to
the condensation of water vapor to form drops of borax solu-
tion was measured in special experiments.  On  the  basis  of
these studies and other system data  and parameters, it was  con-
cluded that the scrubbers could indeed be modified to achieve
the desired result.   The inlet particle size distribution to
the scrubber had a mass median diameter of 14 microns aero-
dynamic and a geometric standard deviation of 5.5.  The par-
ticulate loading was 550 Ibs./hour and the minimum acceptable
scrubber efficiency based on the emission rate limit was 97%.
     Preliminary design and cost estimates were made and on
the basis of these figures, it was decided to build a small
scale pilot plant for further study and confirmation of the
design predictions.   The pilot model was completed and with
the data obtained, more detailed designs and cost estimates
were made.   The results indicated a 20" pressure drop and
400 horse power fan motor would be effective.   This seemed
attractive enough for a go-ahead and U.S.  Borax decided to
proceed with the modification that is now called the A.P.T.
Triple Scrubber.
     It should be noted that the local control agency co-
operated with this program and granted the necessary var-
iances for continued operation while the test work and de-
sign was going on.
     The existing scrubber was a high velocity impingement-
entrainment type in which four annular gas jets impinged on
                            30

-------
the surface of a liquid pool within the scrubber body.  The
gas then passed through a baffled passage which caused several
changes of gas flow direction in order to disengage entrained
liquid.  It was possible to utilize the existing scrubber
body and most of the piping and ducting in the modification.
     Figure 1 illustrates the components of the Triple Scrub-
ber.  First, the process gas stream goes through a saturator
and precleaner section which were added upstream of the ex-
isting scrubber body and where quenching and temperature
reduction occur.  Note that the walls of the circular section
are wetted to prevent material buildup, which is a continual
problem when dealing with near-saturated solutions.  In ad-
dition, sprays bring the gas up to saturated humidity and
wet the particles to foster the particle growth that is im-
portant to the success of the process.
     Second, the gases are conducted through a low velocity
impingement section where larger particles are removed.  Ex-
perience with scrubber systems handling borax solutions and/
or hard water has shown that large deposits of precipitate
can form at various points and can then break loose and de-
posit downstream, where blocking or plugging of apertures
can occur.
     Third, the gases enter a high velocity impingement sec-
tion and are diverted downward through a group of high velo-
city nozzles at the bottom of which are adjustable throats.
These nozzles impart the high velocity that gives the im-
pingement efficiency necessary to remove the finer particles
and attain 97+1 efficiency.  The original annular orifices
were replaced with a set of adjustable throats, which were
placed within the old scrubber body.
     Leaving the high velocity section, the gases travel over
a baffle arrangement and through a mist eliminator (entrain-
ment separator) that is a series of vertical zigzag vanes set
within the scrubber body.   The mist eliminator was designed
                            31

-------
so that it could be removed through an access door for in-
spection and cleaning.
     One scrubber was modified as described above to permit
full-scale prototype testing before modifying the scrubbers
on the other furnace lines.  In addition to the internal
changes, the scrubber body was stiffened by welding on steel
angles so that it could withstand the increased negative pres-
sure.  An old fan which was on hand at the plant was suitable
for the prototype test and it was installed to handle the
increased pressure drop.

Performance
     After construction and start up the modified scrubber
was tested over a range of pressure drop.  The results of
this test series are shown in Figure 2, which shows the re-
lationship between the emission rate and pressure drop.  It
also shows the upper limit for emissions based on the process
weight rule as a horizontal line at 15# per hour.  The curve
shown represents the prediction by A.P.T. on the basis of
results from the small scale pilot plant test.  The points
outlined by squares were determined by the A.P.T. sampling
team shortly after start-up of the full scale test unit and
indicate some degree of instability in the operation result-
ing in some erratic results at mid-range pressure drop.  The
triangles indicate data collected over a longer period of
time by the U.S. Borax stack sampling team and show remark-
able correlation with the predicted operating results.  These
data points are all obtained by E.P.A. method 5 particle
emission tests.
     On the basis of these successful tests on the full scale
prototype, the Company is proceeding to modify its other
scrubbers and has finished the second.  The modified scrub-
bers are being operated at between 20" and 24" w.c. and are
reducing emissions to approximately 9 pounds per hour, as
indicated on Figure 2.
                            32

-------
                      Q
                      i—i

                      cy
CO W
< J
CJ Z
                                                                                                              JD
                                                                                                              ^2
                                                                                                               3
                                                                                                               M
                                                                                                               u
                                                                                                               0)
                                                                                                              •H
                                                                                                              M-l
                                                                                                              •H
                                                                                                               tiC
                                                                                                              •H
                                                                                                              PH
                                                    33

-------
20

19


18

17

16

15


14

13

12
o 10
C/3
a  9
   7

   6

   5

   4

   3

   2


   1

   0
        I    I    I    I    I   I    I    I    I    I    I    I    I
                      LIMIT FOR 11 T/hr FEED
                                          U.S.B,
                           A
                D*-
                                     A.P.T
      PREVIOUS EMISSIONS
      Line#
        1
        2
        3
        4
        5
#/hr
33.6
42.0
61.7
35.8
34.3
      -14" W.C.  AP for above
      I    I    I    I    I   I    I
                  I    I    I    I    I    I
           10
                 20
          30
40
50
60
                 SCRUBBER PRESSURE DROP, in.  W.C.
70
  Figure  2.
          Emission rate vs. scrubber pressure  drop  for
          Triple Scrubber.
                             34

-------
                F/C SCRUBBER DEMONSTRATION ON A
               SECONDARY METALS RECOVERY FURNACE
                              by
                        Seymour Calvert
                         Shamim Gandhi
                Air Pollution Technology, Inc.
                  San Diego, California 92117

                          Dale Harmon
                         Leslie Sparks
             U.S. Environmental Protection Agency
                    Research Triangle Park
                     North Carolina 27711
                           ABSTRACT

     A flux/force condensation scrubbing system was built to
control particulate emissions from a secondary metals recovery
furnace.  Total mass penetration and fractional penetration
measurements were made under several different operating modes.
The performance of the demonstration scrubber was consistent
with the results of previous studies on F/C scrubbing.  The
system was generally capable of 90% to 95% efficiency on par-
ticles with a mass median aerodynamic diameter of 0.75 ymA.
This efficiency was achieved with a 68 cm (27") W.C. gas
phase pressure drop.
                              35

-------
          F/C  SCRUBBER  DEMONSTRATION ON A SECONDARY
                   METALS RECOVERY FURNACE

 INTRODUCTION
     Flux force/condensation (F/C) scrubbing can achieve high
efficiency fine particle collection at lower power input than
the conventional high-energy scrubbing systems.  F/C scrubbing
effects are generally caused by contacting hot humid gas with
colder liquid and/or by injecting steam into saturated gas.
First, the transfer of water vapor toward the cold surface
sweeps particles with it and this effect is referred to as
diffusiophoresis.   In addition, the condensation of water on
the suspended particles causes the particle size to increase.
The resultant particles are easier to collect by inertial im-
paction after growth has occurred by condensation.
     The demonstration scrubber described in this paper was
part of an R§D sequence supported by EPA contracts (see Table 1).
The basic theory and potential applications of F/C scrubbing
were treated in the scrubber systems study and reported in the
"Scrubber Handbook."  The feasibility of F/C scrubbing was
evaluated in the next contract program which covered the de-
velopment of mathematical models for predicting the efficiency
and bench scale laboratory work for experimental verification.
This was followed by two pilot plant studies.  The first study
was a 14 m3/min (500 CFM) multiple sieve plate column with pro-
visions for steam addition and water cooling.  The second study
was a 28 m3/min (1,000 CFM)  horizontal spray scrubber with
water cooling.  The object of the demonstration program as
described in this paper was to test an F/C scrubbing system on
an actual industrial source.
SOURCE INFORMATION
     A secondary metal recovery furnace was selected for con-
trol in the demonstration.  The source operation, which was
ranked among the top 15 particulate sources in the Midwest
Research Institute particulate system study, involved the
burning of insulation from scrap copper wire.
                              36

-------
     The operator loaded metal pallets covered with randomly
coiled scrap wire into a preheated gas-fired furnace.  The
charging door was closed, and as the insulation became hot, it
melted and burned.  Dense black smoke containing hydrochloric
acid, metal chlorides, and other components was emitted through
a breeching into the first of two gas-fired afterburners.  Be-
cause of the batch loading and the variable emissions as the
insulation burned away, the particulate emission rate was very
cyclic.  Opacity was high during the first several minutes and
gradually decreased over the 30-minute operating cycle.
F/C SCRUBBER SYSTEM
     The scrubbing system was composed principally of a cross-
over duct from the furnace stack, a spray type quencher, a
sieve plate column scrubber, and a spray type cooling tower.
The plant as built is shown in Figure 1, while the F/C scrub-
ber system is shown schematically in Figure 2.
     F/C Process Description.  A maximum flow of 200 Am3/min
(7,OOOACFM) of flue gas flowed from the afterburner stack through
a 9.1 m (30') long crossover duct to a spray quencher.  The
gases were preconditioned in the crossover duct and the quencher
by circulating water sprays to reduce the volume and to satur-
ate the gases with moisture.  From the quencher the gas flowed
to a multiple plate column, where it was contacted with a cooled
solution of Na2C03, and the particles were grown and collected
by a combination of diffusiophoresis and inertial impaction.
The gas flowed from the scrubber through an entrainment separa-
tor into the induced draft fan, and then to the exhaust stack.
     Mechanical  Design.   A 9.1 m (30 ') long crossover duct
and associated supporting truss were built from the furnace
stack to the F/C scrubbing system.   Because of the high temper-
ature and the need to conserve gas  enthalpy,  the duct had to be
insulated.   During the program the  carbon steel duct deterior-
ated due to corrosion,  and it was replaced with a type 316 S.S.
                              37

-------
Several spray nozzles were also installed to introduce a Na2C03
solution for cooling and corrosion abatement.
     The quencher included a 1.52 m (5') diameter and 3.7 m (12')
long vertical spray chamber with cocurrent gas-liquid contact
(see Figure 3).  The quencher had a type 316 S.S. shell and also
type 316 S.S. spray nozzles and heater piping.  The external
piping of the quench unit was CPVC (chlorinated polyvinyl chlor-
ide) , and the gas duct to the scrubber was fiberglass reinforced
plastic.  The FRP duct held up well during the program.  A bed
of 2.54 cm (1") polypropylene pall rings was used as an entrain-
ment separator and placed at the bottom of the quencher upstream
of the outlet to minimize entrainment to the plate scrubber.
While only a few percent of the rings were heat deformed during
the four months of operation, this design would not be satis-
factory without positive means for keeping the packing wet.
     The F/C scrubber consisted of a cylindrical 2.3 m (7f6")
and 3.0 m (10') high, five plate structure with provision for
locating a sixth plate in a split tower design (see Figure 4).
The plate scrubber was a hybrid design which offered flexi-
bility in the mixing of sprays and plates, in the number of
plates, and in the points of liquid introduction.  The design
was desirable for the pilot study, but it was much larger and
more complex than would be required for ordinary use.  As
tested, the scrubber unit incorporated several combinations
of plates with 0.48 cm (3/16") and 0.32 cm (1/8") diameter per-
forations.  The scrubber entrainment separator was a tube bank
with ten rows of 1.3 cm (V) diameter tubes in a series.
     Scrubber liquid cooling was done by spraying into the upper
part of a forced draft tower (see Figure 5).  A spray type tower
was used rather than a packed (or filled) tower because of the
concern with solid deposition.  Air was drawn upward through the
tower by a propeller fan, which was located above an entrainment
separator.  Special attention was given to the design and pro-
vision for a high efficiency entrainment separator which was
required because of the fine spray developed in the tower.  The
                             38

-------
unit consisted of three rows of wave plate baffles with spacing
of 1.3 cm  (V) and a wave amplitude of 2.2 cm  (7/8").  Baffle
rows were  offset 0.65 cm  (V) and baffle direction in each row
was also reversed.  The cooling tower piping was PVC, the spray
nozzles were  316 S.S., and the sump tank was polyethylene.
     The induced draft fan for the scrubbing system was a 25 HP,
belt driven blade type, which was capable of 142 m3/min (5,000
SCFM) at 56 cm (22") water column pressure.  In some of the ex-
perimental runs the fan capacity was augmented to give 69 cm
(27") W.C. pressure.
     Instrumentation was provided to measure the parameters
listed below:
     1.  Flow rates
     2.  Temperatures
     3.  Pressures
     4.  Humidities
     5.  Particulate mass loadings
     6.  Particle size distributions
     7.  Gas compositions
The parameters were included to measure both the performance
and the process conditions at different locations of the scrub-
ber system during the test program.
     Liquor System.   Water from the  quencher was drained to a
holding tank and recirculated to the quencher inlet.   Liquor
from the scrubber was collected in another tank and sprayed in
the cooling tower.   The cold outlet  liquid from the cooling
tower was then recirculated to the scrubber.   Liquid disposal
requirements were obtained from the  local authorities and the
demonstration plant was operated in  compliance.  The  major re-
quirements were that the liquid discharge to the sewer be neu-
tralized and that it not interfere with the sewage treatment
plant processes.   The liquor was treated periodically and the
pH was controlled by the addition of soda ash.
                             39

-------
EXPERIMENTAL RESULTS AND DISCUSSION
     The scrubber system was operated for about four months
under several different modes.  The purpose of the testing
phase of the program was to obtain the overall mass efficiency
of the system, the particle collection efficiency across the
F/C scrubber, the operational reliability of the demonstration
plant, and the economics of F/C scrubbing.
     Results.  The F/C demonstration plant was evaluated at
six operating modes as listed in Table 3.  A major division
of the experimental measurements can be made by grouping the
runs before the crossover duct corroded out (Modes A-E) and
those after the new duct was installed (Modes F-J).  This is
because not only was the duct replaced but additional sprays
were installed in the duct, an additional plate was added in
the scrubber unit, and fan pressure capacity was augmented by
the addition of a booster blower.
     The chemical nature of the effluent was determined via
atomic absorption.  The particulate filter catch was largely
composed of metals and halogens while the impinger catch showed
a high concentration of hydrochloric acid (see Tables 4 and 5).
In some runs where fluorinated polymer insulation was present
in the scrap, hydrofluoric acid was encountered.  Because of
the very high halogen acid concentration, we used a basic ad-
dition to the scrubber liquor and corrosion resistant mater-
ials.  For some particle size distribution sampling, cascade
impactors were machined from Teflon.
     Total mass penetration measurements were made for all runs
and fractional penetration measurements were made for selected
runs.  Fractional penetration is penetration related to par-
ticle diameter and it was computed from cascade impactor inlet
and outlet particle size distribution data.
     Overall penetration in an F/C scrubber depends on particle
size distribution, scrubber design, pressure drop, particle
number concentration, particle composition,  and the mechanism

                             40

-------
by which condensation occurs.  Consequently Figure 6, which is
a plot of penetration versus condensation ratio, is subject to
many qualifications.  The condensation ratio is the ratio of
grams of water condensed per gram of dry gas.  In Figure 6, the
two solid curves show the penetration versus condensation re-
sults from a previous pilot plant study on F/C scrubbing, while
the data points show some of the overall penetration data from
the demonstration scrubber.  The penetration decreases with
condensation ratio, and the system performance was better after
the new duct and other modifications were made.  The data points
represented by the triangles were taken while the old duct was
in place, and the data points represented by the circles were
taken after the new duct was installed.
     Fractional penetration characteristics are shown in Fig-
ures 7 to 9.  It is important to note that these are for a
given scrubber design, pressure drop, condensation ratio, and
particle number concentration.  The figures show the experi-
mental penetration curves and that predicted by means of a
design model.  Figure 7 shows a case where the fit between
experiment and prediction is excellent, while Figure 8 shows
a case where the fit was not so good.  Figure 9 shows an aver-
age case where the match between performance and prediction is
good and definitely adequate for design purposes.
     Performance.   The system was generally capable of about
90% to 95% efficiency on particles with a mass median aerody-
namic diameter of 0.75 ymA (about 0.3 urn physical diameter for
particles with a density of 4.0 g/cm3).  This efficiency was
achieved with a 68 cm (27") W.C. gas phase pressure drop.  A
conventional high energy scrubber without F/C effects would
require a pressure drop of roughly 250 cm (98") W.C.  for 90%
particle collection efficiency (a =2.5).
                                 o
     Economics.   In order to compare an F/C scrubber system
with a conventional high energy scrubber, the knowledge from
the demonstration study was used to refine the F/C scrubber

                             41

-------
design.  A cost estimate was made for this design and compared
to a parallel design and cost estimate for a conventional high
energy system.  It is important to note that the high concen-
tration of acid gases required both mass transfer and particle
collection capability.
     The refined F/C system design includes the factors listed
below:
     1.  Scrap - premium #1 wire
     2.  Efficiency required - 88% by mass
     3.  Inlet particle size distribution
          a. mass median diameter = 0.75 ymA
          b. standard deviation = 2.5
     4.  Pressure drop required - 70 cm (28") W.C.
     5.  Equipment
          a. saturator
          b. condenser
          c. venturi
          d. entrainment separator
          e. cooling tower
The control of furnace emissions from the incineration of pre-
mium no. 1 wire would require 88% collection efficiency and a
scrubber pressure drop of about 70 cm (28") of W.C.  High mass
transfer efficiency is required to control the acid gases, so
the total combination of saturator, condenser, scrubber, and
entrainment separator must be included in the design.  A ven-
turi scrubber appears to be the best choice for particle col-
lection after condensation and growth have occurred in the
condenser.  Because the mass transfer capability of a venturi
and entrainment separator is limited, the saturator and con-
denser must be efficient for gas absorption.
     The high energy scrubber system design features are out-
lined as follows:
     1.  Scrap - premium #1 wire
     2.  Efficiency required - 88% by mass
                             42

-------
      3.   Inlet  particle  size
           a.  mass  median diameter  =  0.75  ymA
           b.  standard deviation  =  2.5
      4.   Pressure  drop required  -  238  cm  (94")  W.C.
      5.   Equipment
           a.  saturator
           b.  condenser
           c.  venturi
           d.  entrainment separator
The conventional high energy system without any F/C effects
would also require 88% collection efficiency but a pressure
drop of 238 cm  (94") W.C. for collection of particles with
a mass median aerodynamic diameter of 0.75 ymA and standard
deviation of 2.5  The high energy scrubber would have to in-
clude the same mass transfer capability as the F/C system and
would have the same components as the F/C system except for
the cooling tower and associated apparatus.
     A cost comparison between the F/C scrubber and the high
energy scrubber is given  in Table 6.   The initial capital in-
vestment is lower for the F/C system, and the annual power
cost is also less than the conventional system.  After account
ing for depreciation and  all operating costs including power,
the total annual cost of  the F/C system is about $10,000 less
than the conventional high energy system at this particular
source.  The cost of electric power used was 0.045
-------
  3.   Particle  number concentration  and grown particle sizes
      are  very  important  in F/C scrubbing.
  4.   An improved and simplified design procedure  gave perfor-
      mance  predictions which were  in good  agreement  with ex-
      perimental  results.
  5.   An F/C scrubber capable of controlling no.  1 premium wire
      incineration from this source  would cost $103,000 to in-
      stall  and would have  a total  annual cost of  about $21,000
      based  on  a  ten-year  life depreciation and all operating
      charges.

ACKNOWLEDGEMENT
     The work  described in  this publication  was performed under
Contract No. 68-02-1801 with  the U.S.  Environmental  Protection
Agency.
ABBREVIATIONS AND  SYMBOLS
   C'  =  Cunningham  correction  factor,  dimensionless
   d   =  particle  diameter,  urn
   n   =  particle  number  concentration,  #/cm3
S.S.  =  stainless steel
   a   =  geometric standard  deviation  of particle  size
   o
        distribution
   p   =  particle  density, g/cm3
   ym  =  micrometer  (micron)
                                      !•-           i
 ymA  =  aerodynamic diameter,  d  (C'p  )2,  ym(g/cm3)'5
                              44

-------
PH
PJ
O
PH
H
PH
CO
w
CO
C3
i—i
PQ
Pi
CO
U
PH
43
 rt
10 •
0 o3
rH 0
PH 0
•H (/)
U 3
P! U
•H t/)
rH -H
P.03

00 rH
PI rt
•H -H
43 P
O CH
3 0
P
U
0
•r— i
n
3
CO
0
P
rt
O

uy

O
^rt

P
^
O
p
0
Pi
rH P
O O
t/) PH

U 03
\ Pi
PH rt

1— I
0
1
to

(XJ CXI
r-~
PQ en
PH rH
=t*= * — '

CO
1 — 1
H
iz;

LO
P
U
rt
rH
4_J
Pi
O
u
Cn
l
0
t~-
i
<£,
PH
U
1
XrH*
03 .
£3
p ""?
CO »?
^*
S
0 rX
P O
tfl O
X43
CO 03
Pi
JH rt
0 a:
43
43 rH
3 0
JH 43
U 43
0
rH
-P
•H
H
CO 3
JH
P U
0 CO
&















«*
10
•H
10
X
rH
rt
PI
rt


•>
00
PI

1— 1
0
O3
o
g



£
« o
t/) SH
Pi ^
O
•H rt
p p
u rt
•H O3
O3
(/>
JH
0
43
43
3
rH
O
t/)

U

PH


P
C
rt
rH
P
1
P
0
rH
•H


^
rt
00 JH
PI
•H
43
43
3
JH
U
10

rH 0 rH PnU









































/ — »
(XI

•
rH
0
£>
v — '

P
JH
O
PH
0
Pi

rH
rt
PI
•H
PH

rH
rt
O
•H
P
0
rH
O
0
.r-|
H

to
O
1
to

1
CXI
~\
0
LO
\o

<£,
PH
W

LO
(X)
O
l
(XI
O
i
OO
0

0
u
JH
O
PH

X
3
rH
PH

<-H
O

X
p
•H
rH
•H
43
•H
tfl
rt
0
PH
rt
u
•H
4_j
rt
g
0
j3
p
rt
6

o
to
1
t--

(XI

pa
PH
:*:

CO
1 — 1
f— l
^













OO
PI
•H
43
43
0
JH
U
CO

PI
o
•H
•p
rt

rH
0
p
rt
5-H
0
43
rt
1-1
oo
rH
O
1
LO

1
cxt
*\
O
O
o

<^
PH
P4
(XI
00
o
rH
1
CXi
o
1
oo
"°






^,
0
u
}-H
0
PH

X
^3
rH
PH

MH
O

X
o3
3
P
CO
*\
PH

<4H
O

x
o3
3
P
t/)





^ 	 ^
LO
r^
en
rH
v — '


















00
Pi
•H
43
43
£3
rH
U
CO

Pi
0
•H
P
rt
to
PI
0
03
C
0
u

PH
t/}

o3
PI
rt

0
P
rt
rH
PH •
10
0 3
> P
0 rt
•H JH
to rt
PH
P! PH
•H rt


































to
0
rH
U
•H
-P
JH
rt
PL,

0
PI
•H
PH

<4H
O





P OO
PI PI
rt -H
rH 43
Pn43
1 3
P JH
0 U
rH tO
•H
PnU
^ — _
XPH
JH
O MH
P O
rt
rH X
o ^o
43 3
rt p
i-5 (/)
o
o
fXI
i
^O
r^ ^ 	 ^
1 \O
(X| t~~
*^^ cn
0 rH
0 '— '
\O

<^
PH
W
OO
CXI
to o
rH rH
l =tfc
(XI
O rX
1 (f)
oo rt
vO t~~*

X
rt c
JH O
PH-H
CO P
rt
rH (/)
rt p|
P 0
PI 03
0 PI
isl O
•H U
JH \
O 0
32 U
JH
<4H O
0 PH

X X

3 rH
P PH
CO


X
rt
rH
PH
in

rH
rt
P
PI •
O to
N 3
•H P
JH rt
O rH
45 rt
PH
Pi PH
•H rt











































!-.
0
43
43

rH
O
CO

                                             45

-------
w
w
o
u
w
CO
E-
W
s
w
H
co
>-
CO
w
PQ
PQ
CJ
CO

CJ
o
PL,

CO


H  r<
CO  3
    bo
CO -H
CO U-<
W
u  
(U ctj
+j bO
X RJ •
31
/ — \
p* CO bO
< PA \
IT* ^
H U RS
!2o O
W 0 ,X
co ^ h
CO O ^
^ ^ j*?

/ — >
s ^.s
(i, 'd^s
• ^p =>?
J
O 6 fn
>•< o

S u
Wo
H
•2-
0
1— 1
[—1
1— 1
CO
O
PH
S
O
u
s
w o
H
CO

t~- rH Ol
o to o i
. . . 1
o o o
bo bo bo
v — / v — / v — / <^-
LO
vO rH t-^ CT>
CM rH VO LO
O-J fSJ
CD •<* O O
o oo oo •*
^j- to ^- LO
«j- LO «* to

oo o
00 O rH CM
en CD r-- LO
rH rH


t — • vO ^d"
CTi rH CD CTl
«=j- I~~- LO LO
vO
(D (U
r4 VH
W 3 3
rt -M 4->
O X X ft
•H -H a)
CD S S *->
3 RJ
rH 10 V) ^
PH R$ Ct3
U U

rH CN to -vT




i I I
1 1 1


VO rj- rt
• • •
o cr> cn
<* LO LO

000
o o o
rH rH i-H
<=r •<* ^
to to to

oo oo oo
vO vO vO
LO LO LO


VO ^ ^
CD CD Cn
^J- LO LO





JH f- ft
(D 0) (D
•M 4-> +->
rt RS rt




LO \o r^


                                                 46

-------
  Table 3-   F/C SCRUBBER  DEMONSTRATION PLANT OPERATING MODES
Mode No,
                  Description of System
          5 plates in scrubber, 0.48 cm hole diameter, cold water
          on pass A of scrubber.	
   B
5 plates in scrubber, 0.48 cm hole diameter, cold water
on pass B of scrubber.	
          6 plates in scrubber, 4 plates with 0.48 cm hole dia-
          meter, 2 plates with 0.32 cm hole diameter, cold water
          on pass B of scrubber.	
   D
5 plates in scrubber, 3 plates with 0.48 cm hole dia-
meter, 2 plates with 0.32 cm hole diameter, cold water
on pass B of scrubber.	
          5 plates in scrubber, 3 plates with 0.48 cm hole dia-
          meter, 2 plates with 0.32 cm hole diameter, cold water
          on pass A of scrubber.	
          5 plates in scrubber, 3 plates with 0.48 cm hole dia-
          meter, 2 plates with 0.32 cm hole diameter, cold water
          on pass B of scrubber, sodium carbonate solution spray
          in crossover duct, charge wetted with \\rater.     	
          Sampling at peak of emissions only,  5 plates in scrub-
          ber, 3 plates with 0.48 cm hole diameter,  2 plates  with
          0.32 cm hole diameter, cold water on pass  B of scrubber
          sodium carbonate solution spray in crossover duct,
          charge wetted with watpr.	
   H
Sampling at peak of emissions only,  5 plates in scrub-
ber, 3 plates with 0.48 cm hole diameter,  2 plates with
0.32 cm hole diameter, cold water on pass  B of scrub-
ber, sodium carbonate solution spray in crossover duct,
charge wetted with water,  water spray in crossover duct
          Sampling at peak of emissions  only,  6 plates in scrub-
          ber, 4 plates with 0.48 cm hole diameter,  2 plates
          with 0.32 cm hole diameter,  cold water on  pass  B of
          scrubber, sodium carbonate solution  spray  in crossover
          duct, charge wetted with water, water spray in
          crossover duct.                        	
          6 plates in scrubber,  4 plates  with 0.48  cm hole dia-
          meter,  2 plates with 0.32 cm hole diameter, cold water
          on pass B of scrubber,  sodium carbonate solution spray
          in crossover duct,  charge wetted with water, water
          spray in cross over duct.	
                             47

-------
Table 4.   CHEMICAL ANALYSIS OF PARTICULATES AT
          SECONDARY METALS RECOVERY FURNACE
          ATOMIC ABSORPTION ANALYSIS
       Cu

       Fe

       Na

       Pb

       Zn
11.2%

 8.3%

 6.3%

22.6%

 1.8%
                       48

-------
Table 5.  CHEMICAL ANALYSIS OF IMPINGER SOLUTION AT
          SECONDARY METALS RECOVERY FURNACE
COMPONENT
Cl"
sor
NO;
Zn
Fe
Cu
Pb
Si
Ca
mg/£
6,370
136
93
390
240
85
60
73
17
WT %
85.3
1.8
1.2
5.2
3.2
1.1
0.8
1.0
0.2
                         49

-------
   Table 6.  COST COMPARISON FOR PREMIUM WIRE RECOVERY
COST ITEM*
Total Capital
Investment
Depreciation**
Maintenance
Water
Raw Materials
Power***
Total Annual Cost
COST
FOR F/C
$103,120
10,310
6,185
180
1,650
2,370
$ 20,695
COST FOR
CONVENTIONAL
$111,240
11,125
6,675
180
1,650
11,530
$ 31,160
  * Cost based on 4th quarter 1976
  fe* Based on 10 years straight line depreciation
*** Electric power @ 0.045{/kWh
**
                           50

-------
Figure 1.   Front view of F/C demonstration
           scrubber.
                 51

-------
    u
O
^J
CO

u
       o

       fin
       O
       p
       ex
r*
1
1
i
1
i


u
Z Pi
I-H P4
O O
O E-
O

w


I—I

CO
                                   »
1 LJ
An!
-/ri" 	 	
V1!
1
1
w«
••••



••wv

oi
P4
CQ
Oi
U
cn
                                            w
                                            X
                                            u
                                            z
                                            w
                                            D
                                            ex
D
i	9
                                                                                                      rt
                                                                                                      *
                                            in MH
                                               O
                                            oo
                                            C  C
                                           •H  o
                                           & -H
                                           ,0  -M
                                            3  a
                                            JH -H
                                            O  JH
                                            t/1  U
                                               to
                                           U  
-------
                 HOT STACK  GAS
           /\
ri\\     *i v*      f™
 i\     t\       L
   *7\\    *7^   f"
   /«     /I    l_
                           *     /1
                 ENTRAINMENT

                  SEPARATOR
                                          0
                            GAS TO

                            SCRUBBER
Figure 3.  Quencher unit of F/C demonstration  scrubber,

-------
  GAS
INLET
                                      CLEAN AIR
                       ENTRAINMENT
                         SEPARATOR
                SIEVE
                PLATE
                 #3       GAS
        HH
        3AS t
                 #2
                       GAS
/
                     PASS A
SIEVE
PLATE
 #6
                                           GAS
                                     #4
                   GAS
 J
                                        PASS B
  Figure  4.  Scrubber unit of F/C  demonstration plant,

                            54

-------
                     EXHAUST
          111111II11111111
ENTRAINMENT
SEPARATOR
                                              SPRAY MANIFOLD
                                                AMBIENT
                                                AIR
Figure 5.   Cooling  tower of F/C demonstration scrubber  system.
                            55

-------
    1.0
    0.5
o


t!
rt
*-i
M-l
    0.2
o   0.1
i—i
H
E-
W

g  0.05
   0.01
      0.01
                              4 plates, n =108no./cnr

                                (pilot plant data)
               5 plates, n  =106no./cm

                 (pilot plant data)
                A OLD DUCT - F/C DEMONSTRATION SCRUBBER


                O NEW DUCT - F/C DEMONSTRATION SCRUBBER




               I     I   I  I  I  I I I I	I    III	
                        0.05     0.1

                      CONDENSATION RATIO, g/g
0.5
1.0
       Figure 6.  F/C demonstration scrubber performance
                                56

-------
     0.5
 •P
 O
 52
 O
 W
 ^
 PJ
 OH
 U
 t-H
 H
 Pi
     0.3
     0.2
0.1
    0.05
    0.03
                                        T
                       EXPERIMENTAL-

                       PREDICTED	
  0.3      0.5         1.0        2.0

           PARTICLE DIAMETER, ymA
                                              3.0
Figure 7.  Particle penetration versus  aerodynamic
           diameter for Run 64.
                      57

-------
    0.5
o
•H
4->
L»
rt
O
    0.3
    0.2
H
W
2:
W
PH
   0.05
   0.02
               1   I   I  IT
                          EXPERIMENTAL —

                          PREDICTED	
                                     \
            I    I   I  I  I  I
      0.3     0.5         1.0         2.0

               PARTICLE DIAMETER,  ymA
                                             3.0
Figure 8.  Particle penetration  versus  aerodynamic
           diameter for  Run  74.
                      58

-------
     0.5
U

Cfl
O
I— |

H
2
u;
(X
u
I — I
E-
ai
<
Cu
     0.3
     0.2
     0.1
    0.02
       0.3
             I    I   I   I  I  T
EXPERIMENTAL

PREDICTED — •
0.5        1.0        2.0

PARTICLE DIAMETER, ymA
                  3.0
  Figure 9.   Particle penetration versus aerodynamic
             diameter for Run 69.
                        59

-------
REFERENCES
1.  Calvert, S. ,  J. Goldshmid, D. Leith, and D. Mehta.  Scrubber
    Handbook.  A.P.T., Inc., EPA Contract No. CPA-70-95.  NTIS
    No. PB 213-016.  August, 1972.
2.  Calvert, S.,  J. Goldshmid, D. Leith, and N. Jhaveri. Feasi-
    bility of Flux Force/Condensation Scrubbing for Fine Par-
    ticulate Collection.   A.P.T., Inc., EPA Contract No. 68-02-
    0256.  NTIS No. PB 227-307.  October, 1973.
3.  Calvert, S.,  and N. Jhaveri.  Flux Force/Condensation Scrub-
    bing.  J. Air Pollution Control Association. 24 (10): 947-
    951.  October, 1974.
4.  Calvert, S.,  N. Jhaveri, and T. Huisking.  Study of Flux
    Force/Condensation Scrubbing of Fine Particles.  A.P.T., Inc.
    EPA Contract  No. 68-02-1082.  August, 1975.
5.  Calvert, S.,  and S. Yung.   Study of Horizontal Spray Flux
    Force/Condensation Scrubber.  A.P.T., Inc., EPA Contract No.
    68-02-1328, Task No.  10.  July, 1976.
6.  Handbook of Emissions, Effluents and Control Practices for
    Stationary Particulate Pollution Source.  Midwest Research
    Institute.  Report to NAPCA.  Contract No.  CPA 22-69-104.
    1970.
7.  Peters, M.S., and K.D. Timmerhaus.  Plant Design and Econ-
    omics for Chemical Engineers.  New York. McGraw Hill. 1968.
                             60

-------
    FINE PARTICULATE SCRUBBING - NEW PROBLEMS AND SOLUTIONS
                      Robert W. Mcllvaine
                     The Mcllvaine Company
                   Northbrook, Illinois 60062
                           ABSTRACT

     Because of new processes and legislation, new problems have
arisen that cannot easily be solved by conventional precipitation,
fabric filtration or scrubbing.  A need for new, high efficiency
scrubbing techniques is indicated.
     Several of these problems are described in this paper, and
proposed or actual solutions are discussed.  For example, a foun-
dry shakeout scrubber in Tennessee is in violation because of a
visible blue haze.  The emission is less than .005 gr/SCF.  Add-
ing a wet precipitator is ridiculously expensive, while a fabric
filter replacement is hazardous because of the moisture content.
Some improved scrubbing technique is warranted.  Evaluation of
alternatives is now under way, with the addition of an ionizing
section prior to the scrubber and a wet filter being considered.
     Sinter plants in steel mills produce condensables that are
not captured by fabric filters and precipitators.  Considerable
energy is required with venturi scrubbers to solve this problem .
Results of pilot and full-scale scrubbers are discussed, along
with other metallurgical problems.
                              61

-------
  FINE PARTICULATE SCRUBBING - NEW PROBLEMS AND SOLUTIONS
Many new scrubbing techniques for fine particles are being
considered.  A number of manufacturers Have developed unique
devices for applications where other devices have proved in-
adequate or economically unfeasible.  The Environmental Pro-
tection Agency has undertaken a substantial investigation of
new scrubbing devices for removing particles of less than 1
micron because of their detrimental effect on health.

As a publisher of technical information on the scrubber indus-
try, we have not yet come across any data that has attempted
to summarize the various applications to which the new scrubber
designs may be applied and to match them to the various designs
available.  It is the intent of this paper to provide such a
summary.

Why the need for new scrubbing techniques?

We will not attempt to discuss in this paper the health effects
of fine particulate.  We are starting with the assumption that
a number of companies will be required to remove fine particu-
late from their exhausts in the future.  If more thorough clean-
ing of the air is required and many other types of collectors
exist, why not use something other than scrubbers?  Various
alternatives are presented in Figure 1.  These are dry electro-
static precipitators, wet electrostatic precipitators,  fabric
filters, and incinerators.  There are some applications to
which these pieces of equipment cannot be successfully applied.
By successful we mean that they are neither economical nor
reliable.

Four conditions which increase the cost and/or decrease the
reliability of existing equipment are humidity, corrosion,
high temperature, and high ash resistivity.  Fabric filters
do not function well in highly humid atmospheres.  Corrosion
can be a substantial problem for fabric filters, wet electro-
statics and dry electrostatics.  High temperature creates a
particular problem for fabric filters.  High resistivity ash
makes collection difficult in dry precipitators and is one
reason they are not widely used in industrial applications.
Wet electrostatic precipitators function well because of the
presence of water and are not affected by ash resistivity.
Of course, fabric filters are not affected by ash resistivity.

Again with reference to Figure 1, these devices vary in effi-
ciency versatility.  For example, incineration works well only


                             62

-------
on combustible particles and is not applicable to the majority
of particulate emission problems.  Fabric filters and wet elec-
trostatics are versatile in that they will obtain high effici-
ency on the large majority of applications whereas dry electro-
statics are more limited.

Many of these aspects affect cost.  We must differentiate be-
tween small and large units because the economics are entirely
different for a 10,000 CFM unit than for a 1,000,000 CFM unit.
Dry electrostatics are relatively uneconomical at low volumes.
Incineration is ruled out in many cases because of the high
operating cost, especially when natural gas is used.  In addi-
tion there are applications where incinerators do not have the
required efficiency.  Fabric filters are unpredictable depending
on the humidity, temperature and other physical conditions.  Wet
electrostatic precipitators are extremely costly when handling
corrosive gases.  Dry electrostatic precipitators have a limited
number of applications where they can produce high efficiency
and are not economical for very small installations.  Thus, it
is evident that there are a number of instances when this equip-
ment is not satisfactory and scrubbers should be considered.

What new scrubber types are available for fine particulate
removal?

Before we eliminate the venturi scrubber entirely, we should
note that Venturis are being used at ever-increasing pressure
loss.  For example, there are venturi scrubbers on an iron
foundry cupola operating at 120 inches pressure drop with an
outlet emission of less than .01 grains/SCF.  However, there
are also applications where even at 200 inches pressure drop,
Venturis do riot have acceptable efficiency.

Four types of scrubbers that do have high efficiency potential
are foam scrubbers, electrostatically augmented scrubbers,
steam-assisted scrubbers, and fiber scrubbers.  Figure 2 com-
pares the general performance of each of these types to venturi
scrubbers and conventional low-energy scrubbers.  All of the
new types achieve high efficiency on fine particulate, at least
under some conditions.  The foam scrubber, specifically, can
combine absorption along with particulate collection.  In fact,
one pilot foam scrubber appeared to have the absorption effi-
ciency of a packed tower of nearly 1000 feet high.  The other
types are neither designed for nor do they have high absorption
efficiency.   At present the obvious drawback of these designs
is their relatively high initial cost compared to either the
venturi scrubber or the impingement plate scrubber, although
this initial cost differential can be offset by their lower
energy consumption and operating cost.  The foam scrubber
should be capable of removal of even the very finest particles
with less than twenty inches pressure drop, most of which would
                            63

-------
be in the foam breaker.  Operating cost should be relatively
low if foam is recycled.

On commercial installations, electrostatic augmentation has
proved to have achieved both low energy consumption and opera-
ting cost.  Steam-assisted scrubbers use a considerable amount
of energy; but when used on hot applications, can use waste
heat, thus greatly reducing the net energy consumption and
total operating costs.

Fiber-type scrubbers do require a considerable amount of energy
compared to electrostatically augmented scrubbers, but require
less energy than Venturis.  They are incapable of handling cer-
tain particulate and are therefore potentially unreliable on a
number of applications.  It is really too early to evaluate
the reliability of foam scrubbers, although there is no indi-
cation that they cannot be made to operate quite reliably.

As for liquid requirements, we have not compared liquid-to-air
ratios for the various devices since such a comparison is mean-
ingless.  Liquid is always recycled to some degree, so the more
important issue is how clear the water has to be when it is
introduced to the scrubber and how expensive will the water
clarification system be.  Both steam-assisted scrubbers and
fiber scrubbers require a high degree of clarity.  In the case
of the steam-assisted scrubber the water must pass through a
heat exchanger.  Adaptability to corrosion-resistant construc-
tion is a positive feature of all scrubber types with the ex-
ception of the fiber scrubber and electrostatically-augmented
scrubbers.  However, this statement must be qualified by the
fact that certain manufacturers, such as Ceilcote, have de-
signed an all plastic scrubber which does have good corrosion
resistance.

With the exception of the foam scrubber, there is considerable
operating experience with all the new designs.

What are the specific applications to which new scrubbers are
and will be applied?

This question is a difficult one for two reasons.  First, there
are numerous individual applications, particularly in the chemi-
cal industry upon which one cannot generalize.  Second, unusual
problems arise requiring more elaborate equipment in applica-
tions normally handled by conventional means.  For example,
thousands of scrubbers function adequately on foundry shakeouts.
Emissions are generally less than 0.02 grains/SCF and the stacks
are frequently clear.  However, in at least five U.S. foundries,
there is a substantial visible blue or gray haze being emitted
from the scrubber discharge, and in all five cases local author-
ities have demanded that equipment be improved.
                             64

-------
There are particular coal dryers, sinter plants and foundry
cupolas which are not typical and produce very fine fumes
impossible to control with conventional Venturis.  Therefore,
one substantial area of application of new scrubbers is in
industries and on processes that were ordinarily handled by
other devices.

In addition, there are a number of processes for which there is
no satisfactory alternative. Those applications which are in
this category are quickly determined by examining the various
columns in Figure 3.  Federal S.I.C. numbers (Standard Indus-
trial Classification) for the various applications have been
used for easy reference.  The equivalent venturi pressure drop
requirement tells us immediately whether or not the new scrub-
ber types can compete economically.  On refractory kilns, for
example, a pressure drop of over 100 inches may be required
with a venturi to produce an acceptable plume,  so this becomes
a natural application for a new scrubber type.   Some ferroalloy
applications require pressure drop considerably in excess of
others.  A Fluid-Ionic scrubber has just been applied to a
mineral wool furnace where the pressure drop requirement for
the venturi pilot scrubber was over 100 inches water gauge.

Municipal incinerators are also good applications because of
the high pressure requirements for Venturis.  In fact, there
have been a number of unsatisfactory venturi scrubber instal-
lations on municipal incinerators.

The market size for the various applications in millions of
dollars per year is based on complete installed cost of equip-
ment.  There are at least nine categories which offer a poten-
tial in excess of $5 million per year on an installed basis.
We have not attempted to include coal-fired boilers on the
assumption that the only way this market would even lend it-
self to the new scrubbers would be if standards were issued
which require exceptional control on fine particulate.

The last three columns in Figure 3 indicate the physical as-
pects of the processes described which are most important in
determining whether scrubbers are more attractive than fabric
filtration, precipitators,  etc.  Obviously, for an application
such as liquid waste incineration where high temperature, high
corrosion, and high humidity are all involved,  some form of
scrubber will almost always be used.  In contrast, electric
arc furnaces in the steel industry can be economically con-
trolled by fabric filters or electrostatic precipitators.

High, fair, and low penetration estimates are based on total
market share percentage calculations.  In liquid waste incine-
ration where very high pressure Venturis are required and high
temperature, corrosion, and humidity are involved, the potential
                             65

-------
penetration of new devices is very high.  On»coal dryers where
the pressure requirement on Venturis is relatively low, there
is not much likelihood of a substantial penetration of new de-
vices.  For certain applications, such as black liquor recovery
boilers, there is likely to be some penetration, but other con-
trol devices are well established.

Which companies are involved in development and commercializa-
tion of these new devices?

In addition to the companies listed in Figure 4, several uni-
versities and research organizations are participating through
EPA-sponsored projects.  Our survey includes Asian and European
companies; but because our news-gathering network is more effi-
cient here in the U.S., we may have missed some of the foreign
companies active in this field.

As shown in Figure 4, at least twelve of these companies have
equipment that is now commercially available and at least
thirteen have conducted pilot tests.  Monsanto probably has
the largest number of commercial installations of any in this
group, with over 1500 installations on SC>2 mist removal and
other similar applications.  The foam scrubber lags the
farthest behind with no known commercial installations.  How-
ever, there have been some pilot operations which show a great
deal of promise.

Electrostatically augmented scrubbers have been installed in a
number of different applications.  Both the Ceilcote and TRW
units have been used on installations as large as 200,000 CFM.
Fluid-Ionic Systems has installed several large units.  Heat
Systems-Ultrasonics has specialized in small units often with
soluble fumes.   Chemico has concentrated on coke oven pushing
exhaust and has a complete coke oven pushing car system revolv-
ing around the Aronetics scrubber.  The initial success with
the Lone Star Steel scrubber was on open hearth furnaces so
that steam-assisted scrubbers are now commercially well-
established, but only for a limited number of applications.
For the convenience of readers, names and addresses of com-
panies that are potential information sources in this field
have been included in Figure 5.

In summary, there is a great deal of pilot work that will need
to be done in order to make the new scrubber designs more uni-
versally applicable.  .However, for certain applications, a
number of commercial devices have been established as both
economical and reliable, and it is expected that an increasing
market share will be captured by these devices in the next few
years.
                             66

-------
§
g
Pi
w
m
o
co
z
o
CO
g
o
H
E-i
8
C
•rH CO
4-1 4-1
i 4-1
0 -H
C H
0) -H
•H 4-1
U (d
,c -H 03
Oi M-l )H
•H MH 0)
ffi W >
1 4-1
co co
•H 3
co Q
s£

tjl ^
•H -rH

4-1
•rH
S
0)
a
8
4-1
•H

CO
Q)
O
•D
,__)
rH
Q)
|5

U
O
ffi 4-1

^ rC
CT* Cn
•H -H
tn w



rr^
S Oi
Q -H
1-5 aa





>.
rH
ti rH
0 rH
0 Q)
P4 S




(U
g
4J
(C
a
g
at
EH



rH rH
rH rH
£ £
& £»





£*
O
03
O

J_)
O
U
4J
•H
Ti
•H
§
K


>1 >(
nH nH
r^ ^-J
O O
0 0
04 O4


>— 1
rH H
rH rH
0) OJ
IS 12
g
3
•H
'O
g


|
•rl
^
1





£
a



C"J
Oi
•H
£C






rH
H
(1J
B





>1
rH
O
0
04







>1
i-H
^
O
0
O4

^X(
rH
r)
8
O4


£
Oi
•H




15
a





U
a




jj
a







1








rH
H
1









r-H
rH
0)




H
rH
;s
                                                      C
                                                      O
                                                     •H
                                                      CO
                                                      O
                                                      rl
                                                      M
                                                      O
                                                      U

                                                     4J
                                                      C
                                                     •H
              (1)
              O
             •r)

             a
U
•H
4-1
(3
4J
CO
0
^_4
4J
U
0)
H
0)

fr
Q
U
•H
4J
(d
4J
CO
O
^4
4J
O
0)
rH
0)

-p
0)
s




^j
Q)
4J
rH
•H
t4

U
•H
r4
1






C
o
•H
4J
(0

(U

•H
CJ
C
H
W
0)
•O

•o
•H
0
(d

Jj
a
0)
U
X
w
•-H
                                                 67

-------
CN
H
     CQ
     U
     en
     §
     <


     8
^^
 H






§
rH
w
•H
C


-------













co
§
H
E-i
rt
u
H
ro QJ
S
E> Od
o w
H cQ
u ro
§
u
co
S
g



















a
0
il
W W
o w

H
Q
(H
s
B
H
O
U
B
£
W

§
W
EH

>H A
N W
H 53
w O m
H 1
W J H
3 tj
5! v> v

*. So
> p w
H §5 B
Ot W Oi
H > <










CRIPTION
co
§
U •
H 0
co Z


£ 15 !5 -H -H -H -H CP -H 15 -H
iSSpSlSlSfetuKptilSlij



\
l-^Mh^MKMkCtCkCKK

< \
aajSEKaswKKK


tCSJMEKJWijaS






XXX X



XXX X XXX





o o o o
1 1 1 O 0 0 0 0 0 1 O
r-l
Ul
01
i-H 01
in -H w
M O (U
i
rH O W
•H -r4 >. Q
O 4J M Cn
MO (U C J->
3 > -H (U
iH t3 O Ul rH
0)0) O 01 rH
3 K 01 lD<-l-'-H 0)1-1
u>QO4-> 'SB-H^C!
IU C J< -H H Q4 O

S 5 ti & fi S S co £ £ 5
0 00 0 0 OOOj-l ££
OOOOO^CDCOOOCOCO
OiHrHrHCNOJOJCNCNr^N


tr
33







SB


a






X









o










c
-H
1 Gasificat

S
0
Oi
r*j


•H CT> -H S -H *H O 5 -H -H ^t -H
P^ffifetSfaPMOiSfal^ScM






^
KKC^. rlKKJrHaSErHj


aoBwacawJajxKp






X XX



XXX XX XXXX




o o o
+ O 0 O MO
OOIOIOIOOOI 1
CM CM M ^r







s
18
Sol cn 01 01
id c (1) -P
C3 e> -H O 01 3
(0 r-H [14 ,1^ 18 I8O
S-rl 01 M-l 01 C rH (1)
^jrH 4JM-I3M O i
18 O C O Oj P 01 Qj IB
g>i O 18 b >i 3 43
tlS rHCic OCJCO
p44JrH 2r-rlC[rHE>lI>1
O!8 rlOOrl-r|l8HH
U1I8M OJ 4-'lWOrtJ't3

sassssssssss
8£§HrHrHrHHrHH°8
ojCNfNromnromronmro
mromronrororomnnn


iSS
fa fc "ffi



.1 .1 >n
1— 1 M W


w a »


as a w






X



X X





o o o




01
M
o
01 -P
rl (d
0 M
•P (U
CP 18 C
ti rl -rl
•H 
01
^1 rH 18
•S H*

Sad
§ss
01 CT^ Cft
fO ^C ^











































CTi
s
(0 g
O 5 T3 O>
S 0 | -rl
•HI-HEX

* ^a*



69

-------
                                  FIGURE 4   MANUFACTURERS
FOAM
Monsanto Res-
EPA
Alfa- Laval
Ri ley-En vir
Miura Engr
ELECTROSTATIC
AUGMENTATION
Ceilcote
No Yes 0 Laboratory tests
potential
Yes Yes Several Metallic mining,
show high efficiency
foundries
No Yes 0 Absorption as well as particulate
? ? ? Japanese patent
Yes Yes Many Refractory kilns
issued in 1972
, incinerators, petro-
TRW


Fluid-Ionic


Chemico

Fyltis


RP Ind

APS

WET FILTRATION
Air Industrie


Andersen 2000


Cebeco

Monsanto
Heat Systems-
Ultrasonic
                          leum, coke, calcining kiln

Yes       Yes   Several   Hot grinding, coke oven, scarfing,
                          sludge incinerator

Yes       Yes   Several   Aluminum, steel, pulp S paper, glass,
                          meat smoking, mineral wool cupola

?         ?         ?     (Recent license arrangement with Soc. Lab)

Yes       Yes       -     Aluminum prebake S Soderberg pots, iron
                          S steel, detarring, chemical industries

Yes       Yes       -     pvc textile oven exhaust, foundry cupola

Yes       Yes       -     EPA tests conducted, pilot on hog-fuel
                          boiler

Yes       Yes   Many      Phosphoric & sulfuric acid mist in the
                          chemical industry

Yes       Yes   Many      Phosphorous furnaces, coal dust, acid mist,
                          glass furnaces, galvanizing

Yes       Yes   Many      Ammonium chloride

Yes       Yes   1000's    Acid mist from sulfuric acid plants,
                          fertilizer prill towers, textile driers

Yes       Yes   Many      Aluminum chloride, fertilizer ammoniator,
                          paint fume, oil mist, engine exhaust
                                            70

-------
                                  FIGURE 4   MANUFACTURERS
                                   <>v
                                  nO
                                                            APPLICATIONS
STEAM
ASSISTANCE
Chemico              Yes       Yes   Several   Ferroalloy furnace, coke oven pushing,
                                               carbon black pellet driers

Lone Star            Yes       Yes   Several   Black liquor recovery boilers, open
Steel                                          hearth furnaces

Roskel               Yes       Yes      -      Asphalt plants, incinerators
                                           71

-------
                                       FIGURE 5

                       USEFUL ADDRESSES OF INFORMATION SOURCES
COMPANY NAME AND ADDRESS

Air Industrie
19-21 av. Dubonnet
Courbevoie 92, France

Air Pollution Systems
1114 Andover Park West
Tukwila, Washington  98188
                ACTIVITY
 Manufacturer of wet  sleeve  filter
 Manufacturer  of  electrostatic  ionizers
 and  scrubbers
Air Pollution Technology, Inc.
4901 Morena Boulevard, Suite 402
San Diego, California  92117

Alfa-Laval AB
Pack
S-147 00 Tumba, Sweden

Andersen 2000, Inc.
2000 Sullivan Road
College Park, Georgia  30337

Cebeco Manufacturing Company
1300 North 9th Street
Philadelphia, Pennsylvania  19122

Ceilcote Company
140 Sheldon Road
Berea, Ohio  44017

Chercico Air Pollution Control Company
One Penn Plaza
New York, New York  10001

Environmental Protection Agency
Industrial Environmental Research Lab.
Research Triangle Park
North Carolina  27711

Fluid-Ionic Systems
Div. of Dart Industries Inc.
P.O. Box 20684
2525 East Magnolia
Phoenix, Arizona

Fyltis/Tissmetal Lionel-Dupont
138 Boulevard de La Croix-Rousse
69283 Lyon Cedex 1, France

Heat Systems-Ultrasonics, Inc.
38 East Mall
Plainview, New York  11803

Lone Star Steel Company
2200 West Mockingbird Lane
Dallas, Texas  75235
 Designer of scrubber systems and mist
 eliminators. R§D contractor for E. P.A.
 and other clients.

 Commercial manufacturer of foam scrubbers
 Manufacturer of  CHEAF  unit
 Manufacturer of wet  filter
 Manufacturer of  Ionizing Wet  Scrubber
 Recent  licensee of  Societe  Lab;  Aronetics
 wet  scrubber
 Active  in  the  investigation  of  novel  new
 devices
 Manufacturer  of Hydro-Precipitrol
 Manufacturer  of  electrostatic  liquid
 spray  precipitator
 Manufacturer  of  Mystaire scrubber



 Steam-Hydro Scrubber


72

-------
FIGURE 5  (Continued)

COMPANY NAME AND ADDRESS

The Mcllvaine Company
2970 Maria Avenue
Northbrook, Illinois  60062

Monsanto Enviro-Chem Systems, Inc.
800 North Lindbergh Boulevard
St. Louis, Missouri  63166

Monsanto Research Corporation
Station B, Box 8
Dayton, Ohio  45407

RP Industries
15 Kane Industrial Drive
Hudson, Massachusetts  01749

Riley Environeering, Inc.
7401 N. Hamlin Avenue
Skokie, Illinois  60076

Roskel Air Pollution Control, Inc.
503 Idaho Building
Boise, Idaho  83706

TRW Inc.
One Space Park
Redondo Beach, California  90278

University of Washington
Seattle, Washington  98105
              ACTIVITY

Publisher of manuals, newsletters,
abstracts and search services on all
aspects of scrubbing

Manufacturer of Brink Mist Eliminator
Conducting foam scrubbing tests under
contract to EPA
Manufacturer of Electro-Dynactor
Foam scrubber development
Steam assisted scrubber
Manufacturer of charged droplet
scrubber
Under EPA contract investigating
electrostatic augmentation
                                           73

-------
          ENTRAINMENT SEPARATORS FOR SCRUBBERS
                     Seymour Calvert
                     Shui-Chow Yung
                   Harry F. Barbarika
             Air Pollution Technology, Inc.
                  San Diego, California
                    Leslie E. Sparks
          U.S. Environmental Protection Agency
       Industrial Environmental  Research Laboratory
          Research Triangle Park, North Carolina
                          ABSTRACT

     Experimental results of research and development on
entrainment separators by Air Pollution Technology, Inc.
is presented.  Performance tests were made of four types
of entrainment separators, oriented so that the gas flow
was vertically upward.  The tests showed that re-entrain-
ment occurred at lower superficial gas velocities in this
configuration than when the gas flow was horizontal.
Studies of solids deposition in entrainment separators
were made and an experimental correlation is presented.
Finally, a design method is described which uses the "cut
diameter" versus pressure drop relationship for entrain-
ment separators.
                            75

-------
            ENTRAINMENT SEPARATORS FOR SCRUBBERS

     Gas scrubbers create mist while collecting particles.
This mist is entrained by the gas to an extent depending on
the type of scrubber and the operating conditions.  The en-
trained drops carry dissolved and suspended solids which will
constitute a particulate emission if they are carried out of
the stack.  It is always important and sometimes of over-
whelming importance to separate this entrainment from the gas
before it is emitted to the atmosphere.
     Air Pollution Technology, Inc.  (A.P.T.) has performed
research and development on entrainment separators under
several Environmental Protection Agency (EPA) contracts.  The
work has been reported in the "Scrubber Handbook," Calvert,
et al. (1972) and in contract reports by Calvert  et al.  (1974a)
and (1975) .  A review has also been given in the literature
by Calvert et al.  (1974b).  In this  paper results of further
work concerning entrainment separators are presented.  Three
topics will be covered:
     1. Results of performance tests of four types of entrain-
        ment separators oriented so that the gas flow is
        vertically upward in all cases.
     2. Studies of solids deposition in entrainment separa-
        tors .
     3. A design method for entrainment separators.
PERFORMANCE TESTS
     Four types of entrainment separators were studied  experi-
mentally.  These are described below:
     1. Knitted Mesh - ACS Model 4CA, 10 cm (4 in.) thick;
        0.028 cm (0.011 in.) diameter wires arranged in layers;
        98.21 porosity; and specific surface area of 2.8 cm2/
        cm3  (85.4 ft2/ft3).
                              76

-------
     2.  Packed Bed - Polypropylene pall rings, 2.5 cm  (1 in.)
         in diameter; packed to a 30 cm  (1 ft) height ; and with
         specific surface area of 1.9 cm2/cm3  (58 ft2/ft3).
     3.  Tube Bank - Six rows of 1.9 cm  (0.75  in.) diameter
         tubes, as shown in Figure 1; center-to-center  spacing
         within a row of 3.8 cm (1.5 in.); and center-to-center
         spacing between rows of 2.13 cm (0.84 in.).
     4.  Zigzag Baffles -
         a) Horizontal - Six rows of discontinuous baffles, as
            shown in Figure 2; baffle width of 7.6 cm (3 in.);
            baffle thickness of 0.16 cm  (0.063 in.); space between
            baffles within a row of 7.6 cm (3  in.); space between
            rows of 2.5 cm (1 in.); and angle  to flow of 30°.
         b) Inclined - Same baffle size, spacing and angle to
            flow, as the horizontal.  The baffles were  inclined
            from the horizontal.  Two inclination configurations
            were tested:  30° and 45°.  Figure 3 shows  the 45°
            inclined baffles.

     All these entrainment separators and a cyclone separator
had previously been tested in an arrangement with horizontal gas
flow, as reported by Calvert et al. (1974a,b).  For the latest
series of experiments they were put in a tower with vertically
upward gas flow.  Vertical flow was expected to decrease the
capacity as measured by the reentrainment velocity for  a given
liquid flow rate and not affect primary collection efficiencies.
The entrainment was generated by spray nozzles arranged below
the test section.  Room temperature air and water were used.
     The test section was about 76 cm (2.5 ft) square and the fan
was capable of moving about 1.42 m3/s (3,000 CFM).  The perfor-
mance efficiency was determined from a water balance over the
gas and liquid streams.   The entrainment drop diameters were
averages for specific nozzles operated at a given pressure.
Entrainment liquid flow rate to the spray section was measured by
a calibrated water meter and checked by supply tank level measure-
ments.   The air flow rate was determined by a standard pitot tube

                              77

-------
located at the air inlet.
Performance Parameters
     There are three basic performance parameters.  The first is
the primary efficiency, which is the efficiency at which the
entrainment separator first collects the drops entering it.
When the velocity in the separator becomes so great that the
liquid which collects on surfaces is sheared off, or the liquid
clogs air passages, reentrainment of the collected drops begins
The velocity at which reentrainment first noticeably occurs is
the second parameter.  The final parameter is the pressure drop
through the separator, which will not be discussed here because
flow orientation did not have any effect on the previously repor-
ted (Calvert et al. (1974a,b) pressure drop data.
Performance of Knitted Mesh
     The efficiency of the wire mesh as a function of superfi-
cial gas velocity is shown in Figure 4.  Penetration in percen-
tage is 100 minus the efficiency and is used rather than effi-
ciency for convenience.  Two different liquid flow rates, which
correspond to different entrainment fluxes, were parameters.
Dividing the entrainment flux by the gas velocity will provide
the entrainment liquid to gas (L/G) volume ratio.  Entrainment
L/G's around 1 or 2 £/m3 are in a range which could be expected
from scrubbers.  A set of nozzles providing a drop diameter of
84 pm was used.  Theoretically, a mesh should be 100 percent
efficient once a small minimum velocity is reached; the break
in the curve must represent the onset of reentrainment.
     The onset of reentrainment is related to flooding.  The
predicted flooding velocity, based on correlation by Poppelle
(1958) for an air-water system,is shown in the Figure 4 for the
two entrainment fluxes.  The predicted flooding velocity is
lower than our observed reentrainment velocity. This is because
the onset of flooding does not mean that heavy reentrainment will
immediately occur, since flooding begins at the bottom of the
mesh or packing while reentrainment comes  mainly from the
top of the mesh or packing.
                               78

-------
 Performance of Packed Bed
     Penetration  data for  the packed bed are shown for similar
 entrainment fluxes  in Figure 5.  Here the penetration is not
 zero, but  1  or  2  percent at the lower velocities.  Heavy
 reentrainment begins at about 6 m/s, which is higher than the
 flooding velocity for these liquid flow rates.  The flooding
 velocity prediction is based on Sherwood's correlation for
 dumped packing (reported in Perry and Chilton, 1973).  As with
 the mesh, the flooding correlation underpredicts the reentrain-
 ment velocity.
 Performance of Tube Bank
     Penetration  through the tube bank is shown in Figure 6.
 The 2.5 £/m2-s flux represents a very heavy inlet entrainment,
 such as would come from a venturi.  The dashed line is a pre-
 diction of the primary penetration for 90 ym diameter drops.
 The onset of reentrainment prediction is based on streamlined
 strut data  which  were independent of entrainment flux,
 depending only on the gas and liquid properties.  It was much
 higher than observed.   The experimental onset of reentrainment
 of 3 m/s (10 ft/s) is  the lowest so far shown, but it occurred
 at a very high entrainment flux.
 Performance of Horizontal Baffles
     -
     Penetration of the zigzag baffles in the horizontal  config-
 uration is shown in Figure 7.   Here all the entrainment rates
 are moderate.   The primary efficiency prediction for 90 ym dia-
meter drops is shown as a dashed line.   The onset of reentrain-
ment was predicted, based on the tearing of drops from the baf-
 fle surface,  as being  between 5 and 6 m/s.   The prediction is
 independent of the entrainment flux.   The data,  however,  do show
such a dependence.
Performance of Inclined Baffles
     Inclined  baffles  were used in experiments with large drops
and very heavy entrainment as  shown in Figures 8 and 9.   For the
30° inclined  baffles reentrainment was  noticed at about 6 m/s
gas superficial velocity.   The onset  of reentrainment for 45°
                              79

-------
inclined baffles was not reached.  Figures 7, 8, and 9 graphi-
cally show the effect orientation has on the velocity of the
onset of reentrainment.   The more vertical, in this case the
baffles inclined 45°, are better draining and have definitely
higher reentrainment velocity characteristics.
Performance Test Conclusions
     Primary efficiency can be predicted based on equations
given in the literature and the contract reports.  The present
study showed that orientation of the entrainment separator did
not affect the primary efficiency.  Reentrainment was not that
predictable and predictions based on flooding were too con-
servative.  Orientation definitely had an effect.  Unlike that
used in the present experiments, the configuration with the gas
flowing horizontal has better reentrainment characteristics.

                  SOLIDS DEPOSITION STUDIES
     Possible deposition and removal mechanism  studies, experi-
ments for verification,  and correlations have been made.  All
the deposition mechanisms involve transfer of solid particles
from slurry drops or streams to a solid surface in the separa-
tor.
     The possible mechanisms are:
     1. Gravity settling to non-vertical surfaces.
     2. Inertial impaction due to liquid flow direction changes,
        whether caused by the separator surface curvature or by
        the turbulence of the gas stream.
     3. Diffusion.
     4. Electrophoresis.
     5. Liquid loss from the slurry due to:
        a) Liquid drops falling away from the concentrated
           solids interface.
        b) Evaporation.
        c) Sorption by a dry surface.
Chemical reactions  and precipitation were not included as
part of this study.

                              80

-------
     Once the solid particles deposit on a surface they will
adhere to it for any number of reasons.  Wetting of the sur-
faces seems to be the most important mechanism in overcoming
the adhesion.  Wetting the surfaces is commonly performed by
washing.  Theoretically, the liquid film thickness must be
greater than the slurry particle diameter to provide washing.
     A series of auxiliary experiments was conducted using a
calcium carbonate slurry to provide information on solids de-
position characteristics.  The apparatus for this study con-
sisted of a baffle plate upon which a slurry stream impinged
after being atomized by an air jet.  The major factors affec-
ting deposition and washing were slurry flux, slurry drop size,
and collection surface orientation.
     Based on the auxiliary experiments and a few runs using
slurry in the pilot scale system, the following conclusions
were drawn:
     1. Deposition rate decreases as the slurry flux on the
        surface is increased.
     2, Deposition rate decreases as the liquid film thickness
        is increased.
     3. Deposition rate is higher on an inclined baffle than
        on a vertical baffle due to the increase in settling
        rate of solids suspensions.
     4. Small drops are more susceptible to being caught in
        eddies which would bring them to the back surfaces
        of the baffles.
     5. Small drops which do hit the baffle surface have a
        higher deposition rate than larger drops because
        of their lower localized slurry flux.
     The deposition data were correlated by the following
empirical equation:

                RS = W<|> exp [-(0.13 + 0.534))6]           (1)
                              81

-------
where    Rg = deposition rate of CaC03 on a vertical flat sur-
              face , mg/cm2-s
          W = weight fraction of solids in the slurry
          cj) = slurry flux, mg/cm2-s
          6 = liquid film thickness, ym

     Although slurry flux, "<}>", applies as a multiplicative fac-
tor, the exponential term predominates so that deposition rate
decreases as "" increases.  This equation can be used for
similar slurries and conditions if the slurry flux and the
liquid film thickness are known.  Slurry flux in rows of baf-
fles can be estimated using the efficiency equations.  A pre-
diction method for liquid film thickness is also available in
the report by Calvert et al.   (1975).  The maximum thickness
occurs where "6" is zero, such as at a leading edge.  The row
at which "Rg" first becomes significant would be the one
where washing would be most effective.

                         DESIGN METHOD
     Three steps must be followed to properly select and design
an entrainment separator.  The first step is the determination
of the performance requirements, including collection efficiency,
maximum outlet loading, behavior of emitted entrainment, capa-
city, physical or process limitations, and cost.  The second
step requires gathering of information about the nature of
the liquid entrainment and carrier gas, especially the liquid
drop size distribution.  The final step is the selection of
candidate designs and prediction of the characteristics of
these designs.  For most of the types collection efficiency is
very good as long as the reentrainment velocity is not exceeded.
     For cases where drop collection efficiency requirements
are stringent, such as when a significant amount of the drops
have  diameters   10  urn    or smaller, the prediction of effi-
ciency must be precise.  The "cut diameter method" described
by Calvert (1974) provides a convenient approach to the

                              82

-------
definition of separator efficiency.  For most inertial collec-
tion devices, including entrainment separators, the cut dia-
meter is a power law function of the energy consumed (or pres-
sure drop).   The "cut diameter" is defined as the diameter of
drops collected at 50% efficiency.
     Figure 10 is a plot of performance cut diameter versus gas
pressure drop.  This figure was drawn based on design equations
and experimental correlations.  Shown in Figure 10 are curves
for:  baffles at two angles of attack to the flow direction (not
inclinations to the horizontal),  tube banks with two different
spacings between tubes within a row, packing of one particular
size, and knitted mesh with a specific wire diameter.  The curves
would be different for entrainment separators of other collec-
tion element dimensions but are not much affected by the number
of separator elements.  For instance, while wire diameter changes
move the lines, changing the thickness of the knitted mesh pad
does not (over a reasonable range).
     In any design, one must try to minimize reentrainment by
keeping the superficial gas velocity below the reentrainment
velocity.  Trade-offs on capital  and operating costs are needed
to determine the size and gas velocity or pressure drop of the
entrainment separator.  Since pressure drops are usually low,
capital costs predominate and suggest that a high velocity be
used so that equipment size is kept as small as possible.
     Primary efficiency data, such as presented here for a few
drop sizes,  are also available in the literature for other drop
sizes and conditions.  Predictions can be made based on the
models presented in these reports.  Finally, consideration of
plugging characteristics of the design is needed since solids
will be present on most entrainment separators for scrubbers.
One would probably not choose the mesh or packing designs, which
Figure 10 shows are most power efficient.  This is because
designs which have mostly vertical surfaces that are easily
washed will  be the best to use.
                               83

-------
     In general, it is possible to design entrainment separators
of the types studied so that their performance and pressure drop
are predictable.  It is likely to be more difficult to specify
the entrainment drop size and loading which must be treated.
Reliable measurements of entrainment parameters are required
if one is to make a rational design of an entrainment separator.
                               84

-------
      ooo   o  o  o  o  c
        O  O O O  O  O  O
      oooooooc
     )  O  O O O  O  O  O
      oooooooc
     ^0000000
               GAS FLOW
Figure 1.   Tube bank configurati
on,
                  85

-------
   \\\
   \\\\
   \\\\
     <2> GAS
     I	1 CT r
GAS
FLOW
           ANGLE OF
           ATTACK 30°
Figure 2.  Horizontal baffle configuration,
        86

-------
            INCLINATION
              ANGLE    45
Figure 3.  Inclined baffle configuration

-------
   50
2
w
   40
   30
   20
   10
          I    I    I    I    I
          84 ym DIA.  DROPS

             0.65  £/m2-s
-  Q  0.32 £/m2-s
         	  FLOODING  PREDICTION
          I
        I
I
I
I
I
I
      0123456789   10


      SUPERFICIAL VERTICAL GAS VELOCITY,  m/s


 Figure 4.  Experimental penetration through mesh,

            upward flow.
                    88

-------
     50
 z:
 o
 i—i
 H
 W
 CX
     40
     30
     20
     10
n    i    i    i    i    i    i
 84 urn  DIA.  DROPS

 O 0.78 £/m2-s


 D 0.38 £/m2-s

	  FLOODING  PREDICTION
        012345678    9   10


          SUPERFICIAL VERTICAL GAS VELOCITY, m/s


Figure 5.   Experimental penetration through packed bed,
           upward flow.
                      89

-------
O
I—I
H


U3
  60
  50
  40
  30
  20
  10
                       '90 ym DIA.  DROPS
                          O  2.52 £/m2-s

                          D  0.86 fc/m2-s
                          O  0.43 «Vm2-s

                      	1  PREDICTIONS
      PENETRATION
      FOR dd=90 ym
         I
               I
I
I
I
1
                                 1
I
                                        w
                                        or;
     0123456    789   10

       SUPERFICIAL VERTICAL GAS VELOCITY, m/s

Figure 6.  Experimental penetration through tube bank,
           upward flow.
                  90

-------
o
i— i
H
   50
   40
   30
PS

I  20
   JO
1 1 1 1 1
90 ym DIA. DROPS
O0.70 i/m2-s
DO. 47 £/m2-s
<>0.23 £/m2-s
	 PREDICTIONS
-

"• •
PENETRATION
\
- \
\
\
\ [
- CD OXSO-OKJJ?
i i i i i
i 1 I I



H
Z —
W
i— i
t
2 —
W
W
erf
1
I — •
1 ^D
TJ^O^
i 1 i 1
      0123456789   10


       SUPERFICIAL VERTICAL GAS VELOCITY, m/s



  Figure 7.  Experimental penetration through horizontal

             baffles, upward flow.
                    91

-------
   50
   40
   30
o
I—t
E-
IS  20
2
   10
  I    i    I    i    r
  400  vim  DIA.  DROPS
  ^   3.05  i/m2-s

  O   2.33  £/m2-s

  Q   1.6  £/m2-s

  n   0.98  H/m2-s

	  PREDICTION
           I
      I
I
I
I
      01234   5678    9   10

       SUPERFICIAL VERTICAL  GAS  VELOCITY,   ra/s

Figure 8.  Experimental  penetration  through 30°
           inclined baffles,  upward  flow.
                    92

-------
   50
   40
 «  20
   10
 1,230  urn  DIA.  DROPS
A  2.87  £/m2-s

D  2.38  £/m2-s

O  1.53  £/m2-s


     PREDICTION
          I
    I
I
      01234    56    789   10

        SUPERFICIAL VERTICAL GAS VELOCITY,  m/s

Figure 9.   Experimental  penetration through 45°
           inclined baffles, upward flow.
                   93

-------
                                                                      LO
                                                                           S
                                                                           u
                                                                           a.
                                                                           O
                                                                           Oi
                                                                           Q
                                                                       o

                                                                       o
                                                                       o

                                                                       o
                                                                                 (D
                                                                                 +J
                                                                                 (I)
                                                                                 £
                                                                       •p
                                                                       2
                                                                       (J

                                                                       a>
                                                                       u
                                                                       c
                                                                       rt
                                                                       e
                                                                       ?-
                                                                       o
                                                                       "4-1
                                                                       S-,
                                                                       o
                                                                       ex
                                                                           C/3    CS
                                                                           CO    P<
                                                                           ra    
-------
                         REFERENCES
Calvert, S., J. Goldshmid, D. Leith, and D. Mehta, "Scrubber
     Handbook," EPA-R2-72-118a, NTIS No. PB 213-016, August
     1972.

Calvert, S., "Engineering Design of Fine Particulate Scrubbers
     J. Air Pollution Control Assoc. 24: 929 (1974).

Calvert, S., I.L. Jashnani, S. Yung, and S. Stahlberg, "Entrain-
     ment Separators for Scrubbers--Initial Report, EPA-650/2-
     74-119a, NTIS No. PB 241-189, October  1974a.

Calvert, S., I.L. Jashnani, and S. Yung, "Entrainment Separators
     for Scrubbers", J. Air Pollution Control Association 24:
     971 (1974b).

Calvert, S., S. Yung, and J. Leung, "Entrainment Separators
     for Scrubbers--Final Report", EPA-650/2-74-119b, NTIS
     No. PB 248-050, August  1975.

Houghton, H.G., and W.H. Radford, "Measurements on Eliminators
     and the Development of a New Type for Use at High Gas
     Velocities", Trans. American Inst. of Ch.E. 35^: 427, (1939),

Perry,  R.H., and C.H. Chilton, "Chemical Engineer's Handbook",,
     5th edition, McGraw-Hill, New York, 1973.

Poppelle, E.W., Master's Thesis, Newark College of Engineering,
     1958.
                              95

-------
                 RELATIONSHIPS OF COLLECTION EFFICIENCY
                        AND ENERGY DISSIPATION IN
                          PARTICULATE SCRUBBERS
                                   by

                 Konrad T.  Semrau,  Clyde L.  Witham,  and
                         William W.  Kerlin
                     Stanford Research Institute
                    Menlo Park,  California  94025
                               ABSTRACT
     A pilot-plant-scale investigation was made of how the collection
efficiency of a particulate scrubber is affected by the way that energy
(contacting power) is applied to the gas/liquid contacting process.  The
tests were made with standardized, polydisperse synthetic aerosols of
three different particle sizes.  The performance of a conventional orifice
contactor was taken as the reference for comparison of the other contactors,
which included a contactor with multiple orifices in series, a variety of
pressure spray nozzles, and a combination of an orifice with a pressure
spray nozzle.  Efficiencies were determined over a wide range of operating
conditions with each contactor.  The multiple-orifice series contactor
gave the same performance as the orifice contactor at high contacting power,
but poorer performance in lower ranges of contacting power.  Contacting
power derived from the liquid stream by use of spray nozzles gave lower
efficiencies than were obtained with contacting power derived from the gas
stream in the orifice scrubber.  The deviation between the two levels of
performance became relatively greater with decrease in the aerosol particle
size.  With one exception, the different spray nozzles gave nearly the
same performance at a given contacting power.
                                   97

-------
          RELATIONSHIPS OF COLLECTION EFFICIENCY AND POWER
                 DISSIPATION IN PARTICIPATE SCRUBBERS
     The "contacting power" method for correlation of scrubber efficiency
data was originally developed in a series of four papers(1~^) in which
it was shown that the performance of a scrubber on a given aerosol could
be represented by an expression of the form
                               Nt = apTY                             (1)
where Nt is the number of transfer units, P-p is the contacting power,
and Qi and y are empirical constants characteristic of the aerosol entering
the scrubber.  This is the equation of the "performance curve" for the
scrubber and aerosol.  The contacting power is that which is dissipated
in fluid turbulence, and ultimately as heat, during gas/liquid contacting
in a scrubber.  It may be supplied from the gas stream, the liquid stream,
or a mechanically driven rotor.

     In scrubbers in which all the contacting power is obtained from the
gas stream, the contacting power is essentially equivalent to the gas
pressure drop or (more precisely) to the "effective friction loss" through
the wetted equipment.  An accumulation of data from various sources (•*•"')
as well as a large body of experience have established that, for scrubbers
in this category, the collection efficiency on a given aerosol is almost
wholly dependent on the contacting power and is little affected by the
design or geometry of the scrubber.  In other words, the constants ot and
y of Equation (1) are characteristic of the aerosol, not the scrubber.
Occasional devices or operating modes deviating from this generality
have been ones yielding inferior performance.  Other data (2~^/ indicated
strongly that contacting power derived from the liquid stream or from a
mechanical rotor was probably equivalent to gas-phase contacting power,
but no data obtained under closely controlled conditions have been avail-
able to clearly confirm or disprove this hypothesis.

     In the present investigation, a critical comparison was made of
gas-phase and liquid-phase contacting power.  In addition, the performance
of a multiple-orifice series contactor was studied.  The technique used was
to compare all the gas/liquid contactors over wide ranges of operating
variables, using standardized synthetic test aerosols of three different
                                   98

-------
sizes.  The different devices were compared by reference to the baseline
performance curves established with a conventional orifice contactor.  All
tests were made using air and water at ambient temperature.
Experimental System

     The flowsheet of the experimental scrubber system is shown in Figure 1,
The system could handle up to about 250 m /hr (150 ft^/min) of air
at ambient temperature and pressure, and gas pressure drops across the
scrubber up to about 320 cm WC (125 inches WC).  The scrubber consisted
of a gas/liquid contactor section followed by a cyclone entrainment separa-
tor.  The contactor section consisted of a straight, vertical section of
tubing that could accommodate different types of gas/liquid contactors.

     The air stream was drawn from the room and through the scrubber by
a positive-displacement blower of the Roots type.  The entering air was
metered with an orifice meter, mixed with the test aerosol, and then drawn
through the scrubber.  The reference gas/liquid contactor consisted of an
orifice located in the vertical contacting line, through which the air
and scrubbing water passed downward; this unit was studied in a previous
investigation.(^)  The scrubbing water was fed through a single sidewall
tap located two pipe diameters above the orifice.  The air and water
leaving the contactor were separated in the cyclone, and the air then
passed to the blower.

     The multiple-orifice series contactor (Figure 2) consisted of half-
discs mounted in a staggered array within a tube flanged at one end.  The
assembly was held in position by the orifice flanges that normally held
the orifice plate used as the reference contactor.

     The spray gas/liquid contactors (Figure 3) consisted of five pressure
spray nozzles  (Configurations SSI through SS5) and one combination of a
pressure spray nozzle and an orifice (spray-orifice contactor SO) .  The
SSI nozzle was a specially fabricated concentric tube with multiple per-
forations that discharged solid radial streams of water at right angles to
the direction of air flow.  All the other spray nozzles were commercial
products.  The SS2 nozzle delivered a flat, fan-shaped spray.  Nozzle SS4
produced a hollow-cone spray pattern deflected to a spray angle of 180°.
The SS3, SS5, and SO spray nozzles were all of the solid-cone type,
having nominal spray angles of 100°, 30°, and 60°, respectively.  The
water was supplied to the nozzles at pressures of up to about 13.6 atm
gauge (200 lb/in^ gauge).
                                  99

-------
     Gas flow through the scrubber system was regulated by bypassing air
from the room directly to the blower inlet.  With spray configurations
SS5 and SO, the water jets developed sufficient draft that the scrubber
could be operated without the blower if desired.  In these cases the air
flow through the scrubber was controlled by restricting the discharge of
the air to the atmosphere downstream of the scrubber.

     The scrubber efficiency was determined by sampling the aerosol at the
inlet and outlet of the scrubber.  At the outlet, the aerosol was collected
on a Nuclepore membrane filter.  At the inlet, the sample was collected
in a miniscrubber composed of four Greenburg-Smith inpingers in series
followed by a Nuclepore filter; this unit both characterized the aerosol
and determined its quantity.
Aerosol Generation

     The aerosol generation system, shown schematically in Figure 1, was
similar to that used in a previous investigation.(^)  A. cloud of droplets
of the solution of aerosol material (ammonium fluorescein) was swept out
of the aerosol generator by a small metered flow of saturated cold air,
then dried by mixing with an additional metered flow of dry air.  The
resulting dry aerosol particles were then mixed with the main air stream
passing to the scrubber.

     The aerosol generator' ' was a greatly improved development of a
generator used in the previous investigation.(^)  The basic element was
a commercial Mist-02~Gen ultrasonic nebulizer, but this was combined with
other components designed to aid production of a stable and consistent
aerosol.  The particle size of the solid aerosol was determined by the
initial droplet size and by the concentration of ammonium fluorescein in
the solution.  The ammonium fluorescein aerosols were spherical, nonhygro-
scopic, and only very slightly soluble in water; small samples could be
determined precisely by fluorometric analysis.  Three polydisperse aero-
sols of similar character but different particle sizes were produced by
varying the solution concentrations; their size characteristics are pre-
sented in Table I.

     The aerosol concentrations at the inlet ranged from about 0.02 to
20 mg/std m^, depending on the air flow rate and the aerosol being generated,
This low aerosol concentration range was beneficial in limiting the tendency
of the particles to flocculate.
                                   100

-------
Results and Discussion

     Although the aerosol generator was reasonably stable, the particle-
size characteristics of the aerosols sometimes underwent substantial varia-
tions, and the possibilities of greater variations could not be ignored.
An accurate, rapid, and simple method was needed to determine the variabil-
ity of the aerosol, both for initial study of the generator itself and for
monitoring the aerosol during scrubber testing.  The miniscrubber was used
for this purpose as well as for determining the inlet concentration to the
scrubber.  It was always operated at the same conditions, with a total
pressure drop across the four impingers of 200 cm WC.  The quantities of
aerosol collected by each impinger and the final filter were determined
individually, permitting determination of the efficiency of each stage and
of the complete train of impingers.  Plotting the cumulative number of
transfer units against the corresponding cumulative pressure drop produced
a performance curve for each miniscrubber run (Figure 4).  Since the
miniscrubber was always operated in an identical manner, this performance
curve was characteristic of a given aerosol and was therefore termed the
"signature curve" or "signature" of the aerosol.  Although the signature
gave no absolute measurement of the aerosol particle size, it was a very
sensitive indicator of changes in the particle size and/or size distribu-
tion.  If the signature of the inlet aerosol for a given scrubber test
varied from the norm by an unreasonable margin, the test could be dis-
counted.  Figure 4 shows the average signature curves for Aerosols D, E,
and G as well as that for Aerosol F, which was originally intended to
serve as the finest of the aerosols but was found unsatisfactory because
of contamination of the solution.  The contaminant was apparently silica
dissolved in the ammonium hydroxide used to prepare the solution.  Its
effect was to cause formation of nonspherical particles.

     Performance curves for the orifice scrubber were determined for each
of the three aerosols (Figures 5, 6, and 7).  For convenience, the con-
tacting power was expressed as effective friction loss.  In each case, a
two-branched performance curve was  clearly defined, with  the  lower branch
having a  slope greater  than 1.0.  These phenomena were  first  observed,
but much  less clearly,  in  the earlier  study with a  similar ammonium
fluorescein aerosol.(5)  It is speculated that  this type  of behavior may
be associated with a transition between controlling mechanisms in the
collection of the particle.

     The performance of the multiple-orifice series contactor on these
aerosols is shown in Figures 8, 9,  and 10, in which the baseline perform-
ance curves for the orifice scrubber are shown  for reference.  With
Aerosols D and E, the data fitted the  correlations for  the orifice scrubber
in the upper range of effective friction loss, but fell below the orifice
scrubber curves in the lower ranges (Figures 8 and 9).  With Aerosol G,
                                   101

-------
the performance of the multiple-orifice series contactor was inferior to
that of the orifice contactor except at the highest levels of effective
friction loss that were attainable with the experimental system (Figure 10)
The trends of performance with the three aerosols show that the deviation
of the multiple-orifice contactor performance from that of the orifice
scrubber increased with a decrease in aerosol particle size.  There is
no obvious explanation for the performance deviation and, indeed, no
reason to assume that such a deviation should necessarily be characteris-
tic of all forms of series contactors or scrubbers in series.

     In the tests with spray contactors, there was always a gas pressure
drop (or friction loss) across the scrubber, so that there were no cases
in which all the contacting power was derived from the liquid stream.  The
gas-phase and liquid-phase contacting powers and the total contacting
                                           •3
power were computed in units of kWh/1000 m  from the following equations:
                            PG = 0.02724 FE                          (2)

                            PL = 0.02815 pf (L/G)                    (3)

                            PT = PG + PL                             (4)

where
                 Fg = effective friction loss, cm WC

                 P£ = liquid feed pressure, atm gauge

                  G = gas flow rate, m-^/min

                  L = liquid flow rate, liters/min.
     Some of the tests of the SS5 spray configuration and the spray-
orifice contactor were made without use of the blower, but with the draft
supplied by the water jets.  The operating mode was thus the same as
that of the ejector venturi scrubbers that are in wide use.  In all these
cases, there was a net gas pressure rise, rather than a pressure drop,
across the scrubber, so that the nominal PQ value was negative.  The
nominal P  value did not, in fact, constitute contacting power since it
         O
represented energy that was converted to gas pressure instead of dissipated
                                   102

-------
in  turbulence.  The energy of  the  liquid jet was used in  two ways, partly
as  contacting power and partly to  pump the gas through the scrubber.
Furthermore, a portion of the  part of the energy that was used  to pump
the gas was eventually converted to gas-phase contacting  power.  Of the
total gas pressure head developed  by the liquid jet, part was lost by
friction in passage through  the scrubber; neither of these two  quantities
was measured explicitly in the experiments, but only the difference be-
tween them, which was the net  pressure rise across the scrubber.

     The tests of the spray  scrubbers covered wide ranges of gas pressure
drops, liquid and gas rates, and feed water pressures.  When the efficiency
data were correlated as functions  of the total contacting power, the per-
formance varied widely but was always definitely inferior to the perform-
ance of the orifice scrubber when  a substantial fraction  of the total con-
tacting power was derived from the liquid.  However, the deviation between
the performances of the orifice scrubber and the spray scrubbers was found
to be correlated with a simple parameter _f, the ratio of  the liquid-phase
contacting power to the total  contacting power.  These correlations for
tests with Aerosol D are illustrated in Figures 11, 12, and 13.  Complete
data are presented in the final project report.

     Since efficiency data were not determined for precisely equal values
of _f, they were grouped according  to J: values lying in short ranges.
For _f values in the range 0-0.39,  the data points fell on or close to
the orifice scrubber performance curve (Figure 11).  At least part of
the scatter in the data was due to the wide range of _f values.  At f_
values in the range 0.90-0.99, three correlation curves appeared, one
for configurations SS5 and SO, one for SSI, SS2, and SS3, and one for
SS4.  The performance of SS4 was much inferior to that of the other
configurations, for undetermined reasons.  However, at least part of
the difference between the performance of SS5 and SO and  that of SSI,
SS2, and SS3 has an apparent explanation.  Because the liquid jets of the
SS5 and SO contactors supplied part of the propulsion for the gas stream,
part of the contacting power nominally calculated as liquid-phase was
actually gas-phase, as expained above.  Consequently, the actual value of
j? was lower than indicated.  If this factor could be determined quanti-
tatively, the performances of  the  SS5 and SO contactors and of  the SSI,
SS2, and SS3 contactors would  certainly be in closer agreement.

     The performances of the SS5 and SO contactors were identical within
the precision of measurement (Figures 12 and 13).  Adding the orifice as
a gas-phase turbulence promoter in conjunction with the spray nozzle
afforded no interaction increasing the effectiveness of the liquid-phase
contacting power.  The performance of the two contactors when operated
in the "ejector mode" (Figure  13) was nearly the same as that obtained
at _f = 0.90-0.99 with the blower in operation (Figure 12).


                                   103

-------
      In  Figures  12  and  13,  the  performance  curves  for  the  spray  scrubbers
were  arbitrarily drawn  parallel to  the  branches  of the reference perform-
ance  curves  of  the  orifice  scrubber,  since  inspection  of all  such data
obtained (°) made this  procedure seem as  reasonable as any other.

                                                                    (8)
      Additional  tests of  spray  scrubbers  were made  using Aerosol G.
The efficiency  data showed  the  same trends  as were  found with Aerosol  D,
but the  relative deviation  between  the  performances  of the  orifice  and
the spray scrubbers was greater.(8)   Thus,  the relative inferiority of the
spray scrubbers  was a function  of the aerosol particle size and  increased
with  decrease in the particle size.   If the deviation  in performance
consistently decreases  with increase  in particle size, then for  collection
of particulates  larger  than perhaps 2-3 pm, liquid-phase contacting
power may not be substantially  less effective than gas-phase contacting
power.  If so,  this phenomenon  would  explain earlier observations  (2-4)
that  liquid-phase and gas-phase contacting  power appeared  to be  fully
equivalent.  Most of these  early observations were  made in  relation
to collection of dusts  that contained substantial  fractions of relatively
coarse particles.

      At  least part  of the apparent  systematic inferiority  of liquid-
phase contacting power  may  result from  an unavoidable  approximation in
the calculation of  PL  from  Equation  (3).   The calculated value  of  P-^  is
equivalent to the theoretical kinetic energy of  the  liquid  spray  emerging
from  the nozzle. It is assumed that  all  this energy is effective in gas/
liquid contacting and hence can be  considered as contacting power.  In
fact, the energy delivered  in the spray is  reduced  by  friction in the
nozzle,  which is not known  and  would  be difficult  to determine.  The
internal construction of  the SS4 nozzle indicates  that the  friction loss
should have  been greater  than that  in the other  nozzles, which may  ex-
plain at least  some of  the  relative inferiority  of  the SS4  performance.

      Once the spray has been discharged from the nozzle, some of  the
kinetic  energy may  be lost  by dissipation in ways  that do not contribute
to gas/liquid contacting.   It has been  speculated  that inelastic  impact
of spray on  the  pipe walls  might be sufficient to  account  for the apparent
relative inferiority of liquid-phase  contacting  power. However,  this
conjecture does  not appear  to be entirely consistent with  the relative
performance  of  spray nozzle arrangements  that differed widely with  re-
spect to spray  impact on  the walls.

      It is notable  that all  the  spray contactors except SS4 gave  about
the same performance at a given  contacting power despite radical differ-
ences  in the nozzle  designs  and  arrangements.  Certainly,  the various
nozzles delivered sprays with different drop sizes  as  well  as different
configurations.  Yet, such differences  in the nature of the sprays
evidently had little or no  influence, independent of the contacting power,
on the collection efficiency.

                                   104

-------
Nomenclature




f    = ratio PL/PT


FE   = effective friction  loss, cm WC


G    = volumetric gas  flow rate, m-Vmin


L    = volumetric liquid flow rate,  liters/min


Nt   = number of transfer  units, dimensionless


     = ln[l/a - T7)]


pr   = liquid feed pressure, atrn gauge

                                             o
P£   = gas-phase contacting power, kWh/1000 m

                                                3
P    = liquid-phase contacting power, kWh/1000 m
 JLi

                                         3
P    = total contacting power, kWh/1000 m


O.    = coefficient of  PT in Equation (1), [kWh/lCOO m3


^    = exponent of P~,  in Equation (1), dimensionless


77    = fractional collection efficiency, dimensionless
                                  105

-------
             Table I   Test aerosol characteristics
Code
designation
D
E
G
Particle-size
rangea
(Mm)
0.2 -5
0.15-2.6
0.07-1.3
Mass -median
diameter
(pm)
1.05
0.68
0.42
Geometric
standard
deviation0
1.58
1.69
--
o
 Observed from electron micrographs.

 Aerodynamic diameter determined by cascade impactor.
£
 Particle-size distribution taken as  log-normal.
                             106

-------
 Rooi
 Air
                       Aerosol Sweep
                       Air
Compressed
   Air
                                          Inlet
                                         Sample
                    MANOMETERS


am
\ir —
ow









L_

( Tl)


AEROSOL DRIER
AND
\ /

L
— CT0
-®
IT?
A '
<^j
1


ROTAM
                                                      GAS/LIQUID
                                                      CONTACTOR
                                                       SECTION
                                                             I      I
                                         PUMP
                                                  Fresh
                                                  Water
                                                                                                  By - Pass
                                                                                     Outlet
                                                                                     Sample
                                                                       CYCLONE
                                                                   ' ENTRAINMENT
                                                                      SEPARATOR
                Water
                 To
                Dram
                                                   To Stack
                                                                                          BLOWER
                                                                                                         SA-4380-50
                       Figure 1   Flowsheet  of  experimental scrubber system
                                              GAS
                                             FLOW
                             SEMICIRCULAR /
                                  ORIFICE
                                              r
                                           h
                                               i
                                           j
3__	 WATER
     INJECTION
                            Figure 2  Multiple-orifice series gas/liquid contactor
                                             107

-------
                                                                        _ WATER
                                                                          INJECTION
 SS1
                                            SS2
                                                                                               SS3
                                             30°	\
SS4
                                            SS5
                                                                                                SO
                                        Figure 3   Spray gas/liquid contactors
                                                       108

-------
                  I      I
                                                                AEROSOL  0     —
Z  10
                                 PRESSURE DROP — cm WC
            Figure 4   Minrscrubber signature curves for aerosols D,  E, F, and G
      1000




SA-438O-1G
                 Figure 5   Onfice scrubber performance curve  for aerosol D
                                        109

-------
1      I     I    I   I   I  I  !  I            I       I    1    I   I   I  I  L
 Figure  6  Orifice scrubber performance curve for aerosol E
                           100
                          TION  LOSS — em WC
      1000

SA-43BO-16
 Figure 7   Orifice scrubber performance curve for aerosol  G
                        110

-------

i—i—r
     o

                                 ORIFICE
                                SCRUBBER
                                                            Mill}
          EFFECTIVE FRICTION LOSS — cm WC
                                                                     1000


                                                               SA-4380-17
Figure 8   Multiple-orifice series scrubber  performance on aerosol  D
                    oHiFice      r~
                    SCRUBBER   X
                    Xo
                                     P'

                                100
               EFFECTIVE FRICTION  LOSS — cm  WC
                                                                 SA-4380-18
 Figure 9   Multiple-orifice series scrubber performance on aerosol E
                             111

-------
                                                           I       I     I    I   MIL
£  10
                                                           I       I     I    I   I   I  I  I
                              EFFECTIVE  FRICTION LOSS — cm WC
              Figure 10   Multiple-orifice series scrubber  performance on aerosol G
      1000




SA-4380-20
                              CONTACTING  POWER  — kWh/1000 m3
                                                                             SA-4380-33
         Figure 11   Performance of spray scrubbers on aerosol D at  values of  f = 0-0.39
                                          112

-------
                          CONTACTING POWER   kWh/1000 r
                                                                      SA-4380-38
Figure 12   Performance of spray scrubbers on aerosol D at values of  f = 0 90 - 0 99
                                           ORIFICE
                                          SCRUBBER
              1       I    I    I   I  I  I I
                                       1 0

                        CONTACTING POWER — kWh/1000 m3
  Figure 13  Performance of spray scrubbers of ejector configuration of aerosol D
                                   113

-------
 References

 1.    C. E.  Lapple  and  H.  J.  Kamack,  "Performance  of wet  dust  scrubbers,"
      Chem.  Eng.  Progr.  51 (3):110 (1955).

 2.    K. T.  Semrau,  C.  W.  Marynowski,  K.  E.  Lunde,  and C.  E. Lapple,
      "Influence  of power  input  on efficiency  of dust scrubbers," Ind.
      Eng. Chem.  50 (11): 1615 (1958).
      	          r*~t

 3.    K. T.  Semrau,  "Correlation of dust  scrubber  efficiency," J. Air
      Pollut.  Contr.  Assn. 10 (3):200 (June  1960).
                           *~>-l

 4.    K. T.  Semrau,  "Dust  scrubber design--a critique on  the state of the
      art,"  J.  Air  Pollut. Contr.  Assn.  13  (12):587 (1963).

 5.    K. T.  Semrau  and  C.  L.  Witham,  "Wet scrubber liquid utilization,"
      EPA-650/2-74-108,  U.S.  Environmental  Protection Agency,  Washington,
      D.C.  (October 1974).

6.    A.  B.  Walker, "Enhanced scrubbing  of  black liquor boiler fume by
      electrostatic preagglomeration:  A pilot plant study," J.  Air
      Pollut.  Contr. Assn. 13 (12):622 (December 1963).
      —   , -  , —	 __ _   - T   I -_. ^^

 7.    A.  B.  Walker  and  R.  M.  Hall, "Operating  experience  with  a flooded-
      disc scrubber--a  new variable throat  orifice contactor," J. Air
      Pollut.  Contr. Assn. 18 (5):319  (May 1968).
                 	"----•--- ^j

 8.    K.  T.  Semrau, C.  L.  Witham, and W.  W.  Kerlin, "Energy utilization
      by wet scrubbers,"  Final report,  Contract No. 68-02-2103, U.S.
      Environmental Protection Agency, Washington,  D.C. (Draft in review,
      March 1977).
                                    114

-------
       DIFFUSIOPHORETIC PARTICLE COLLECTION UNDER TURBULENT CONDITIONS

                       Peter J. Whitmore and Axel Msisen
                       Department of Chemical Engineering
                       The University of British Columbia
                       Vancouver, B.C.,  Canada  V6T 1W5.

                                  ABSTRACT
     Diff usiophoresis, which is the movement of aerosol particles induced
by diffusing gases, was studied under conditions of turbulent flow.  A novel
theory for diffusiophoretic particle removal is presented which depends solely
on the non-steady-state forms of the continuity equations for the particles
and gas mixture.  It is therefore independent of equipment geometry and patterns
of mass transfer or gas flow.  Collection efficiencies are predicted for
particles moving with the Schmitt and Waldmann velocity  (or small particle
velocity) as well as the local mean mass and local mean molar velocities of
the gas mixture.  It is shown that, in the latter two cases, the particle
collection efficiency equals the fractional mass and molar removal of gas,
respectively.


     Experimental studies were performed by passing a binary gas mixture
containing aerosol particles up through a wetted wall column (0.0254 m I.D.
and 0.77 m in length), countercurrent to a flow of water.  One component was
inert, while the other was partially absorbed.  The resulting particle removal
was determined by measuring the inlet and outlet aerosol number concentrations
in the gas.  The soluble gases tested were ammonia and trimethylamine, while
the insoluble gases were helium, methane, nitrogen, argon and freon 12
(dichlorodifluoroniethane).  Uniform latex particles of 0.50, 0.79, 1.011, 2.02,
and 5.7 microns in diameter were used.  The experimental results were in
satisfactory agreement with theoretical predictions.
                                     115

-------
      DIFFUSIOPHORETIC PARTICLE COLLECTION UNDER TURBULENT CONDITIONS
                               INTRODUCTION
     Small particles suspended in diffusing fluid mixtures experience a
force and move in a preferred direction.  Although this phenomenon, called
diffusiophoresis, was first reported by Aitken   ' last century it did not
receive much attention until recently.  The diffusiophoretic velocity was
found to be almost independent of particle size, which is in marked contrast
to Brcwnian diffusion and inertial deposition; diffusiophoresis can therefore
provide a useful technique for separating micron-size particles from gases .
One simple way of achieving particle collection by diffusiophoresis is to add
a gaseous component to a particle bearing gas and then remove the component in
an absorber.  Transfer of the gaseous component causes the simultaneous
deposition of particles by diffusiophoresis in the absorption medium.  If the
gaseous component is a vapour, such as steam, and its removal is effected by
condensation, the particle collection can also be enhanced by thermophoresis
                    (3  4)
and particle growth   '   .  All of these effects usually occur in conjunction
with Brcwnian motion and inertial deposition.
     The diffusiophoretic velocity depends on whether the particles are "small",
"similar" or "large" compared with the mean free path of the surrounding
molecules.  The corresponding particles are usually classified as belonging to
the "free molecule", "transition" and "continuum regimes".  Basic theories
describing diffusiophoresis in gases at rest have been extensively reviewed
elsewhere       and need not be repeated here.


     A model for diffusiophoretic removal of particles from turbulent gases
                         (8)
was first proposed by us     and a similar, but less accurate, model was later
                              (9)
published by Azamiouch et al.   .  A brief critique of the latter paper as
t Diffusiophoresis also occurs in liquid systems but very little is presently
  known about it (2) .
                                    116

-------
well as an extension of our original work to diffusio- and thermophoretic


                                                         (2)
collection of particles of all sizes will appear shortly    .  Our model was '



based on the assumption that mass and particle transfer take place from a



turbulent fluid core across a laminar film adjacent to the transfer surface.



This approach is not quite correct since disturbances from the turbulent core



prevent the establishment of a laminar, steady-state film.  In spite of this



shortcoming the film model was expected to yield good approximations since



it has proven to be very successful in classical heat and mass transfer studies.




Although attempts have been made to predict diffusiophoretic  particle removal



from laminar gas streams (eg., by Meisen et al.    ) , no general models have



so far been developed.








     The theory which we present in this paper is of a different genre from



all previous ones.  It is based on the particle and fluid continuity equations,



written in special forms, and it does not depend on the assumption of a laminar,



steady-state film.








     The efficacy of diffusiophoresis for removing micron-size particles



from gas streams in laminar flow was demonstrated experimentally by Schmitt     ,



Goldsmith and May (12), Stinchcombe and Goldsmith (13), and Meisen et al.^.



In order to minimize equipment size (or cost), industrial applications usually



require that particle collectors operate under turbulent conditions.  The first



experimental study of diffusiophoresis under turbulent conditions was performed



by Whitmore and Meisen     who dealt with the removal of 0.79 ym latex particles



from nitrogen-ammonia mixtures in a wetted wall column.  Water was used as the



absorption medium.  The results gave satisfactory agreement with the predictions



of the film theory derived for the case where the particles adopt the local mean




                                    117

-------
mass velocity of the gas mixture.  The experiments were later repeated in


                                      (2)
more detail using improved techniques     but the results were essentially



the same.
                      (9)
     Azarniouch et al.    have also published data on diffusiophoresis under



turbulent conditions.  However, since diffusiophoresis was not the only,  and



in many cases not even the dominant, mechanism for particle removal, their



results are hard to interpret.  Similar difficulties arise with the papers of


              (3  4)
Calvert et al.  '    who studied diffusiophoresis in the presence of conden-



sation effects which led to particle enlargement.
                                 THEORY




     We consider the situation where a particle-bearing gas mixture  is



passed through a device which removes at least one of the mixture components



partially or completely by contact with a separate phase.  This process is



accompanied by particle transport towards the interface which arises from bulk



fluid flow as well as diffusiophoretic particle motion relative to the bulk



flow.   (The existence of bulk flow towards the interface is most obvious in the



case of a single component, absorbable gas).  The present theory is applicable



to situations with and without bulk flow towards the interface and the following



main assumptions are made:
t  Our theory also holds for suspensions of particles in liquids, and might

   have application in problems such as particle transfer during liquid-liquid

   extraction.
                                    118

-------
          1.  Particles move relative to the surrounding gas mixture
              only by diffusiophoresis.

          2.  The volume fraction of particles in the gas mixture is
              small.

          3.  Particles which reach the interface are not re-entrained.
The first assumption excludes the possibility of significant particle

collection by Brownian motion, inertial mechanisms as other effects.  The

second assumption is needed since the particle and fluid continuity equations,

on which our theory is based, are only valid for systems with low particle

concentrations.  Many industrial aerosols fall into this category.



     The particle continuity equation may be written as




          |£   +   V • n v      =   0                                (1)
where n is the particle number concentration, t is time and v  is the particle

velocity due to diffusiophoresis as well as bulk fluid flow.  The general

continuity equation for the fluid is given by
                  _  .  c' v1  =  0                                  (2)



where c1 denotes the fluid concentration expressed in some units (as yet

unspecified) and v1 is the appropriate related fluid velocity.



     Equations (1) and (2) are quite similar and they may be combined by
                                    119

-------
choosing a fluid velocity, v1, which is identical to the particle velocity,

v    This is, however, only possible if an appropriate fluid concentration,
-P
c', exists.  Assuming this to be the case gives
          3(n/cl)  +  v1 • V_   (n/c1)  =  0
             O t
                                                     (3)
for c' 7* 0.  Equation  (3) can be written as
where
          D(n/c')  =
            Dt
           D
           Dt
=  0
                                                    (4)
  +  v' •  v
(5)
 The quantity D/Dt denotes the time derivative for an element following the

 fluid as it moves with velocity v1.  When v1 is the mean mass velocity, v,

 equation  (6) defines the well-known substantive derivative.  It follows from

 equation  (4) that n/c1, the number of particles per unit of fluid, is constant

 along the path of the element.




      The particle collection efficiency for an arbitrary piece of equipment

 operating at steady-state is defined in the usual way.
n v' dA -
A. j
n v1
A .
in J out
dA] /

n •<
A.
•'"in
                                                                      (6)
 where A.  and A   denote the inlet and outlet cross-sectional areas, respectively.
                                      120

-------
The component of v" perpendicular to the elemental area dA is represented



by v'.  Similarly, the fraction of fluid removed (in appropriate units) is



given by
          ef  "
c' v1 dA -
A.
in
c' v1 dA] /
Aout
c1 v1 dA
Ain
                                                                        (7)
Since we are now dealing with steady-state conditions, the variables on the



right hand side of equations (6) and (7) must be interpreted as time-averaged



rather than instantaneous values.








     If n/c1 is constant throughout the inlet fluid, i.e. the entering gas is



well mixed, then n/c1 must also be constant everywhere in the equipment.  It



follows immediately that
          £P  =
                                                             (8)
Thus, when the inlet fluid is well mixed, the fractional particle collection



equals the fraction of incoming fluid removed, measured in the appropriate



units.  This is the primary result of our theory.








     In many practical situations, one component of the fluid is essentially



inert, and passes through the equipment unchanged.  In this case the previous



result can be readily expressed in terms of inlet and outlet fluid concentrations.



Rewriting the expression for particle removal efficiency by using appropriate



average values of c1 and v1 for the terminal conditions, yields
ep  =
                    -  [c' V
(9)
                                   121

-------
Also, if component 1 is inert, it follows that
           [ c- V A]in =  [cj v' A]out                                  (10)
where c,1 is the concentration of component 1 in the mixture.  Substitution



of equation  (10) into  (9) gives :
           £p = 1 -  [ cVcj]out / [ c'/cil^                            (11)






We can now obtain specific results for various particle velocity expressions.



Four different cases will be considered.
1.  Large Particles Adopting the Fluid Mean Mass Velocity





     As discussed previously, we consider this to be the most likely velocity



adopted by large particles, i.e. particles falling into the continuum regime.



 (This  is also the probable result for particles in liquids) .  Equations  (1)



and  (2) can be combined by putting
          c1 = p and c,1 = p,





where  p, denotes the mass of component 1 per unit volume of mixture.  From



equations  (7) and  (8) it follows that the fraction of particles removed equals



the fraction of fluid mass  transferred.  If component 1 passes through the



equipment unaffected, equation  (11) gives
                [p/pl]out  /  [p/pl]in
                                     122

-------
For gas mixtures it is often more convenient to work in terms of molar units
and equation  (12) can be re-written as
where the mean molecular weight is given by
                 k
          M  =  C   Yi Mi
The mole fraction and molecular weight of component i are denoted by y-
and M. , respectively.
2.   Particles Adopting the Fluid Mean Molar Velocity
     Derjaguin et al.   '     obtained this result theoretically for the
velocity of large particles in gases.  Although we have criticized their work
elsewhere     and the agreement with experimental measurements is not very good
as shown later in this paper, the mean molar velocity may nevertheless
constitute a useful limiting case for the diffusiophoretic particle velocity.
As before, equations (1) and (2) can be combined by putting
          v   =  v*  =  v1 ,
          c'  =  c  and  c,1 = c.
where v* denotes the mean rnolar velocity.  The molar concentration and
partial molar concentration of component 1 are represented by c and c,,
                                    123

-------
respectively.  Equations (7) and (8)  therefore indicate that the fractional



particle removal is equal to the fraction of fluid moles transferred.   When



component 1 is inert/ equation (11)  gives
                                                                         (15)
or
3.  Small Particles





     Derjaguin and Bakanov      first derived an expression for small particles



in gases by using methods of molecular mechanics.  Etnploying similar techniques,



but allowing for more complex interactions between the gas molecules and



particle surface, Waldmann independently obtained a comparable equation.  Later



derivations by Bakanov and Derjaguin     , Mason and Chapman     ,  Derjaguin and



Yalamov     and Annis et al.     suggested the presence of small second-order



terms and other minor changes in Waldmann's expression.  However, none of these



modifications affect the particle velocity significantly in most practical



situations.
     Although previous workers mainly considered diffusiophoresis in binary



gases, the general velocity expression for systems consisting of k components



can be deduced:
                     ai
                                    124

-------
where N. denotes the molar flux of component i and a. is a coefficient



depending on the interaction between component i and the particle surface.
     There is no standard form of the fluid continuity equation written in



terms of this velocity but such an equation can be constructed.  In order to



do this we define a new unit of concentration which may be called the "root



mass concentration".








          c.  = a. i/M. c.                                                  /-i o\
           i    111                                                 (18)






Other related variables can then be defined to form a compatible system







          N^  =  a. ^M. N.                                               (19a)





                  k



          cr  =  Y^,   °r                                                (19b)


                 1=1





                  k                 k                 k

           V-     \	*   T-      V-    r  *     	         V	V

          v  =   l^  W  /  c  =  2_ ai /M. N± / c 2_, ai ^ j±       (20)


                 1=1   X           1=1               1=1




The velocities given by equations (17) and (20) are now identical.  The fluid



continuity equation in terms of root mass units can be derived from the standard



continuity equation for component i:








              +  V • N. =  0                                            (21)
                      ..
Multiplying this equation by a. &. and summing over all component yields
                                    125

-------
         1-     V^  a.  M.  c.   +  V .V^  a.  /M.   N.  = 0                (22)
                L—/   111     —  L—i   1   1  —1
or, upon using the definitions (18)  to (20),
              +  v  •  cr vr  =  0                                        (23)
                 —
By putting
          Yp
           i       r    j  i       r
          c1   =  c   and c!   =  c.
the conditions for equations (7) and (8) are satisfied.  Hence, the fraction


of particles collected equals the fraction of root mass units of gas transferred



from the fluid.  If one component in the gas mixture is inert, equation (11)


becomes

                                                   Ei  \

                           a.  ^ Yi / Y]_ ]Qut / [  £  a± M± Yi/Y;L]in   (24)
     The coefficients a. are usually of similar magnit\jde and can therefore be



cancelled from equation (24) without great loss of accuracy.  Under these



conditions the particle collection efficiency falls between the mass and molar


fraction of gas removed.






4.  Particles of Intermediate Size




     There is no satisfactory theoretical expression for the velocity of



                                    126

-------
particles in the transition regime.  However, sane experimental evidence



suggests that the velocity falls between that for small and large particles



The following approximation therefore suggests itself for the transition regime
          v   .      =  f  v      ,, +   (1-f)  v   ..
          -p, trans       -p, small           -p, large
                                                   (25)
where f is a constant of order 1 or smaller.  Although the particle collection


                                                                           (8)
efficiency can be developed for this expression by the earlier film model    ,



we are unable to make the same derivation using the continuity equations.



Indeed, the only type of expression which seems to permit the latter approach is
    Erv

a. Ma  N. / c       a. M
 11-1
                                               rv


                                            .   a
                                      I—,   i  i
where a is an exponent probably lying between 0.5 and 1.  However, equation



 (26) has no theoretical basis.  By defining the concentration of component i



as a. Ma c., it is easy to find the particle collection efficiency.  If component



1 is inert, the result is
                             i Mi *i / TL W / f   E ai Mi VVdn  (27)
                            EXPERIMENTAL TORK
     A simplified diagram of the experimental equipment is shown in Fig. 1,



The apparatus and procedure were essentially those described earlier    .



                                   127

-------
However, the Royco Optical counter (Model 225) and sampling technique were



improved to further reduce errors in particle count.  These modifications are


                            (19)
fully described by Whitmore
     Most of the experiments were performed with a wetted wall column 0.77 m



in length and 25.4 mm in diameter.  Latex particles 0.50, 0.79, 1.011, 2.02



and 5.7 vim in diameter were used and the following gas systems were investigated:



helium-ammonia, methane-ammonia, nitrogen-amttonia, argon-ammonia, freon 12



(dichlorodifluorcmethane) - ammonia, nitrogen-trdmethylamine.  Water was the



absorption medium in all cases.







     The Reynolds number for the gas phase ranged from 2000 to 8000 with respect



to the column wall and, as in previous experiments, no particle removal by



mechanisms other than diffusiophoresis could be detected.  A check was also



made for possible end-effects by performing some runs with a column of identical



design except for its shorter length. As expected from the  theory,  there were



no end effects.
                                    128

-------
                               RESULTS






     The main experimental data are shown in Figs. 2  to 7 .  Three theoretical



values for the particle removal efficiency are also given in each figure.



These were calculated from the expressions developed in the Theory section



and correspond to the assumptions that the particles travel respectively with



the local gas mean mass velocity, the mean molar velocity, or the velocity



predicted for small particles.  For some gas mixtures experiments were made



with particles of different sizes.  In these cases (Figs.  4, 6 ,  7) only the



average particle removal efficiency is shown for each particle size and set of



experimental conditions.  Where particles of just one size were used (Figs. 2,



3,5) all results are plotted.








     The experiments for each gas mixture, except helium-ammonia, were con-



ducted with a constant flow rate of inert gas and a range of inlet flow rates



of transferred gas.  The former flow was chosen so that all results fell into



the turbulent regime.  In the helium-ammonia experiments helium useage was



minimized by operating with a constant inlet flow rate of ammonia and varying



the helium flow rate.








     The figures show that the experimental data tend to lie within the



envelope formed by the efficiency predictions of the mean mass and mean molar



velocity models.  Furthermore, the trend exhibited by the set of data for each



specific gas system, particle size, and type of experiment, is very similar to



the trends of the three theoretical predictions.  Since the experiments cover



such a wide range of mass transfer conditions and gas physical properties,



such agreement provides strong evidence that the theory of particle deposition




                                   129

-------
presented in the previous section is correct, at least as a first order



approximation.







     At the same time, the position of the data relative to the three models



varies with both gas system and particle size, and the data as a whole do not



correlate well around the predictions of any specific model.  As discussed


           (19)
by Whitmore    , the possible systematic errors alone could account for the



lack of correlation with the mean mass velocity model.  This does not mean,



however, that the position of the data relative to the various models is



without significance.  It is unlikely that the variation in data position is



caused primarily by systematic errors, as these would tend to nullify the



observed similarity between the trends shown by each set of data and the



corresponding model predictions.








     The only obvious remaining explanation for the experimental results is



that the particles were exhibiting transition regime behaviour.  According to


          (7)
our theory    large particles move with the mean mass velocity of the fluid,



while small particles travel at essentially the Schmitt and Waldmann (or small



particle) velocity.  The latter falls generally between the mean mass and mean



molar velocities and the removal efficiencies in the transition regime should



therefore lie between the same limits.  Furthermore, for a given gas system



and particle size, these efficiencies should show the same trend as the model



predictions, when experimental conditions are systematically varied.  Since



the outlined behaviour is identical to that exhibited by the experimental



results obtained in the present study, it is concluded that the particles



generally fell within the transition regime.  The Knudsen number, i.e. the



ratio of the mean free path of the gas molecules to the particle radius, ranged



from 0.015 to 0.30.

                                      130

-------
Previous work    on phoretic effects indicates that transition behaviour is



to be expected for such Knudsen numbers.
     The transition regime behaviour precludes definite confirmation that large



particles move at the mean mass velocity.  There is, however, an indication



that they do, since for small Knudsen numbers the data tend to correlate around



the corresponding model predictions.  In contrast, only a small fraction of the



data agree  with the predictions of the mean molar velocity model.  This suggests



that the theories of Derjaguin et al.   '    , which predicted this velocity for



large particles, may be incorrect.








     Practical and cost aspects were considered by Whitmore^  '  and will be



published separately.




                                CONCLUSIONS
     A new theory has been presented for predicting diffusiophoretic particle



collection in mass transfer equipment operating under turbulent conditions.  The



theory is general and depends primarily on the particle and fluid continuity



equations as well as the fundamental diffusiophoretic particle velocity.  The



experimental results, which agree quite well with the theoretical predictions,



generally fall between the limits given by particles adopting the mean mass and



mean molar velocity of the fluid.  This suggests that the particles used in this



study exhibited transition regime behaviour.





                              ACKNOWLEDGEMENTS




      The financial support provided by the National Research Council of



 Canada, the Government of the Province of British Columbia and the University



 of British Columbia are gratefully acknowledged.




                                      131

-------
Figure 1.  Schematic diagram of experimental  equipment,
                           132

-------
   o
   UJ
   O
   LL
   LL
   LU
   o

   UJ
   OC
   LU
   -I
   O

   DC
                           THEORETICAL  MODEL:
                           MEAN MASS
                           MEAN MOLAR
                           SMALL PARTICLE
Figure 2.
                 0.2      0.4     0.6      0.8     1.0
              MOLE FRACTION OF INERT AT INLET
Results  for helium-ammonia mixtures with 0.79 micron
diameter particles.  Inlet flow rate of ammonia held
constant at 6.0 x lO'^m

               133

-------
     0.6
 O
 UJ

 O
 u_
 UJ
I
UJ
oc
HI
d
     0.5
o  0.4
     0.3
 O  0.2
     0.1
     0.0
                             A
        L THEORETICAL MODEL:

          MEAN MASS
          MEAN MOLAR      - —

          SMALL PARTICLE  - —
       0.0      0.2      0.4     0.6      0.8      1.0

           MOLE FRACTION OF INERT AT INLET
Figure 3.  Results for methane-ammonia mixtures with 0.79 micron

         diameter particles.   Flow rate of methane held constant
         at 5.98 x lO'" m3/sec.
                        134

-------
     0.6
  O
  UJ
  O
  E
  LL
  LU

  I
  UJ
  DC
  UJ
  _J
  O
     0.5
0.4
0.3
0.2
     0.1
     0.0
                          PARTICLE DIAMETER,
                               MICRONS:
      THEORETICAL MODEL:

      MEAN MASS      —
      MEAN MOLAR    -——
      SMALL  PARTICLE	
        0.0     0.2     0.4      0.6      0.8     1.0
            MOLE  FRACTION  OF INERT AT INLET
Figure 4.  Results for nitrogen-ammonia mixtures with particles
         of various diameters.  Flow rate of nitrogen held
         constant at 6.0 x 10"'*m3/sec.

                       135

-------
       0.6
    0.5

O

|  °'4
u_
u.
UJ

I  0.3
   UJ
   cc
   UJ
   O
   cc
    0.2
       0.1
       0.0
                         \
             THEORETICAL MODEL:
             MEAN MASS
             MEAN MOLAR
             SMALL PARTICLE	
Figure 5.
      0.0      0.2      0.4     0.6      0.8     1.0
           MOLE FRACTION OF INERT AT INLET
      Results for argon-ammonia mixtures with 0.79  micron
      diameter particles.   Flow rate of argon held  constant
      at 5.6 x 10~ltm3/sec.
                      136

-------
       1.0
  I
  UJ
  cc
  UJ
  _i
  O
  cc
       0.8
  LLJ
  O
  iZ   0.6
  u.
  UJ
0.4
       0.2
       0.0
               \
                  »
                  \
                    PARTICLE DIAMETER,
                         MICRONS:
                         A   0.79
                         O   1.01
                         D   2.02
                    THEORETICAL MODEL
                    MEAN MASS
                    MEAN MOLAR	
                    SMALL PARTICLE —
         0.0     0.2      0.4      0.6     0.8      1.0
             MOLE  FRACTION  OF INERT AT INLET
Figure 6.  Results for freon 12 (dichlorodifluoromethane) -ammonia
         mixtures with particles of various diameters.  Flow rate
         of freon 12 held constant at  0.98 x ICT1* m3/sec.
                        137

-------
      0.3
   O
   UJ
   O
u   0.2
   Li.
   UJ
   LU
   QC

   UJ

   O

   QC
    0.1
      0.0
                           PARTICLE  DIAMETER,
                                MICRONS:
                                V    0.50
                                A    0.79
                                O    1.01
                                D    2.02
                                0    5.7
             \
 .0
                 o\
                    k
                      "<
^
       o
                              .$
           THEORETICAL MODEL:  \  \
           MEAN MASS      •       \\
           MEAN MOLAR	\ V
           SMALL PARTICLE	
Figure 7.
       .6         .7        .8        .9       1.

         MOLE FRACTION OF INERT AT INLET

      Results for nitrogen-trimethylamine mixtures with par-
      ticles of various diameters.  Flow rate of nitrogen
      held constant at 6.0 x 10"'*m3/sec.

                    138

-------
                                REFERENCES
 1.  J. Aitken, "On the Formation of Small Clear Spaces in Dusty Air".
     Trans. Roy. Soc. Edin., 32: 239, 1883.

 2.  P.J. Whitmore and A. Meisen, "Estimation of Thermo - and Diffusiophoretic
     Particle Deposition".  To be published in Can. J. Chem. Eng.

 3.  S. Calvert and N.C. Jhaveri, "Flux Force/Condensation Scrubbing".
     J. Air Poll. Control Assoc., 24: 946, 1974.

 4.  S. Calvert, J. Goldshmid, D. Leith and N.C. Jhaveri, Feasibility of Flux
     Force/Condensation Scrubbing for Fine Particulate Collection, E.P.A. -
     650/2-73-036, Oct. 1973.

 5.  L. Waldmann and K.H. Schmitt, "Thermophoresis and Diffusiophoresis
     of Aerosols" in Aerosol Science, C.N. Davies  (ed), Academic Press, London,
     1966, p. 137.

 6.  B.V. Derjaguin and Yu. I. Yalamov, "The Theory of Thermophoresis and
     Diffusiophoresis of Aerosol Particles and their Experimental Testing"
     in Topics in Current Aerosol Research (Part 2), G.M. Hidy and J.R. Brock
     (eds), Pergamon Press, Oxford, 1972, p. 1.

 7.  P.J. Whitmore and A. Meisen, "The Theory of Diffusiophoresis for Large
     Aerosol Particles".  J. Aeros. Sci., ]_: 297, 1976.

 8.  P.J. Whitmore and A. Meisen, "Diffusiophoresis under Turbulent Conditions".
     J. Aeros. Sci., 4_: 435, 1973.

 9.  M.K. Azarniouch, E.J. Farkas, N.E. Cooke and A.J. Bobkowicz, "Removal of
     Micron-Size Particles from Gases by Turbulent Deposition and Diffusiophoresis".
     Can. J. Chem. Eng., 53_: 278, 1975.

10.  A. Meisen, A.J. Bobkowicz, N.E. Cooke and E.J. Farkas, "The Separation of
     Micron-Size Particles from Air by Diffusiophoresis".  Can. J. Chem. Eng.,
     4_9: 449, 1971.

11.  K.H. Schmitt, "Quantitative Untersuchungen zur Staubabscheidung in
     Diffundierendem Wasserdampf".  Staub, 21-  173, 1961.

12.  P. Goldsmith and F.G. May, "Diffusiophoresis and Thermophoresis in Water
     Vapour Systems" in Aerosol Science, C.N. Davies  (ed), Acadanic Press, London,
     1966, p. 163-

13.  R.A. Stinchcombe and P. Goldsmith, "Removal of Iodine from the Atmosphere
     by Condensing Steam".  J. Nucl. Eng. 20; 261, 1966.

14.  B.V. Derjaguin, Yu. I. Yalamov and A.I. Storozhilova, "Diffusiophoresis of
     Large Aerosol Particles".  J. Coll. and Interf. Sci., 22_: 117, 1966.

15.  B.V. Derjaguin and S.P. Bakanov, "The Theory of the Motion of Small Aerosol
     Particles in a Diffusion Field".  Dokl. Akad. Nauk. SSSR, 117; 959, 1957.

                                     139

-------
16.  S.P. Bakanov and B.V.  Derjaguin,  "The Motion of a Small Particle in a
     Non-Uniform Gas Mixture".  Disc. Farad. Soc., 30;  130, 1960.

17.  E.A. Mason and S. Chapman, "Motion of Small  Suspended Particles in Non-
     uniform Gases".  J. Chem.  Phys.,  36;  627,  1962.

18.  B.K. Annis, A.P. Malinauskas and  E.A. Mason, "Theory of Diffusiophoresis
     of Spherical Aerosol Particles and of Drag in  a Gas Mixture".  J. Aeros.
     Sci., <4: 271, 1973.

19.  P.J. Whitmore, "Diffusiophoresis  under Turbulent  Conditions".  Ph.D. thesis,
     Dept. of Chem. Eng., Univ.  of British Columbia,  Vancouver, Canada, 1976.
                                    140

-------
        IMPROVED DESIGN METHOD FOR F/C SCRUBBING
                     Seymour Calvert
                    Shamim A. Gandhi
             Air Pollution Technology, Inc
               San Diego, California 92117
                        ABSTRACT

     A new design method for predicting the efficiency
of a flux force/condensation (F/C) scrubber is presented,
This method uses mathematical models which define the
particle collection due to diffusiophoresis,  the amount
of particle growth due to water vapor condensation, and
the collection efficiency for the grown particles.   Pre-
dictions for a sieve-plate column are confirmed by pilot
demonstration plant data.
                          141

-------
          IMPROVED DESIGN METHOD FOR F/C SCRUBBING

     Flux force/condensation (F/C) scrubbing provides a means
for the enhancement of fine particle collection efficiency
through the effects of condensing water vapor.   Several phenomena
are simultaneously involved and a detailed mathematical model is
complex.  In a series of previous studies l '2 '3 ' **  we have
developed a design method (i.e., performance prediction) based
on a model which required the simultaneous solution of several
differental equations.  While it produced useful  results, it
was a time-consuming procedure, even with the aid of an elec-
tronic computer.
     In the course of refining our design method  we arrived at
the conclusion that the flux force effects could  be treated
separately from the condensation effects in many  scrubber situa-
tions.  While the condensation-induced improvement in inertial
impaction efficiency could be handled conveniently, the flux
force deposition prediction remained cumbersome.   Recently the
senior author became aware of a concept developed by Whitmore5
which provided the key to simplifying the prediction of the flux
force effects in F/C scrubbing.  By incorporating Whitmore's
concept into our model we have developed a much simpler design
method which is convenient to use.  This revised  model will be
described below.
BASIC CONCEPTS - Before proceeding to the details of the mathe-
matical model the basic concepts and outline of the approach
will be discussed.  If we consider a typical F/C  scrubbing system,
it might have the features shown in Figure 1.  The gas leaving
the source is hot and has a water vapor content which depends on
the source process.  The first step is to saturate the gas by
quenching it with water.  This will cause no condensation if the
                              142

-------
particles are insoluble, but will if they are soluble.  There
will be a diffusiophoretic force directed away from the liquid
surface.
     Condensation is required in order to have diffusiophoretic
deposition, any growth on insoluble particles, and extensive
growth on soluble particles.  Contacting with cold water or a
cold solid surface is done next to cause condensation.  While
condensation occurs there will be fliffusiophoretic and thermo-
phoretic  deposition as well as some inertial impaction (and,
perhaps,  Brownian diffusion).  The particles in the gas leaving
the condenser will have grown in mass due to the layer of water
they carry.
     Subsequent scrubbing of the gas will result in more particle
collection by inertial impaction.  This will be more efficient
than impaction before particle growth because of the greater
inertia of the particles.  There may be additional condensation,
depending on water and gas temperatures, and its effects can be
accounted for as discussed above.
     One  can apply this general outline of F/C scrubbing to a
variety of scrubber types and Figure 2 shows a multi-plate
F/C scrubber system.  It can be seen that the gas is saturated
before entering the plate column, although this is not always
necessary.  The first plate can serve as the saturator and par-
tial condenser.  Generally, the efficiency of heat and mass
transfer  is so high (say, 80%+) on a well designed plate that
most of the condensation occurs on the first plate.
     In subsequent plates the gas is scrubbed by inertial impac-
tion and  there will be a minor amount of additional condensation.
We have shown a simple counter-current column but other varia-
tions are possible.
     The  mathematical model is based on the process just des-
cribed for a plate-type F/C scrubber.  It is outlined on the
following page.
                             143

-------
        I.   Saturate the gas before plate  1
             A.   Particles are collected at size  "d   ".
                                                  pl
             B.   No condensation occurs  on  the  particles
                 (they are assumed insoluble).
       II.   Contact on plate 1
             A.   Particles are collected by impaction in  the
                 bubble formation zone,  still at  size "d   ".
             B.   Condensation occurs and particles grow to
                 "V-
             C.   Diffusiophoretic deposition  removes  some
                 particles from the gas  in  the  froth  layer
                 on the plate.  Thermophoresis  and centrifugal
                 deposition are neglected.
      III.   Contact on plate 2
             A.   Particles are collected by impaction in  the
                 bubble formation zone at size  "d ".
                                                 p2
             B.   Negligible condensation occurs.
       IV.   Contact on subsequent plates has same characteristics
             as  plate 2.
       Although the model is somewhat idealized,  it  is  well  within
the bounds of engineering accuracy and the precision of experi-
mental measurements.  One can always revise any stage of the
model with a closer approximation of parameters if he wishes
to examine the sensitivity of the method to parametric values.
In our evaluations of the model we found that further refinements
did not produce significant changes in predictions.
                              144

-------
DIFFUSIOPHORETIC DEPOSITION
     Particle deposition by dif fusiophoresis was described by
the following equation in our previous models 2 ' 3 ' ** :

                           D
                          (i-y)  vM2 (i-y)  \dr
     UPD = c' DG

where  D- = diffusivity of water vapor in carrier gas, cm2/s
       MI = molecular weight of water, g/mol
       Ma = molecular weight of non-transferring gas, g/mol
        y = mole fraction water vapor, dimensionless
        r = distance in the direction of diffusion, cm
     The molecular weight and composition function represented
by "C i", describes the effect of molecular weight gradient on the
deposition velocity corresponding to the net motion of the gas
due to diffusion (the "sweep velocity").  For water mole fraction
in air ranging  from 0.1 to 0.5, "Ci", varies from 0.8 to 0.88.
We used a rough average of 0.85 for "Cj", for computing "u ~"
and consequent particle collection efficiency by integrating
over the period of condensation.
     Whitmore5  concludes that the fraction of particles removed
from the gas by diffusiophoresis is equal to either the mass
fraction or the mole fraction condensing, depending on what
theory is used for deposition velocity.  In other words, it
is not necessary to follow the detailed course of the conden-
sation process, computing instantaneous values of deposition
velocity, and integrating over the entire time to compute the
fraction of particles collected.  One can simply observe that
if some fraction of the gas is transferred to the liquid phase
it will carry along its load of suspended particles.
                              145

-------
     We have used Whitmore's general concept but with two modifi-
cations.  First, one can see from equation (1) that Whitmore's
theory would be comparable to assuming that the particles move
with the same velocity as the gas phase.  We have chosen to
retain   the correction for molecular weight gradient, which
means that we will compute the particle collection efficiency
as 851 of the volume fraction of gas condensing on the cold sur-
face .
     The second modification concerns how to compute the proper
value of the volume fraction of gas condensing.  The problem is
that not all of the condensate goes to the heat transfer surface;
some of it goes to the suspended particles.  As will be shown in
detail later, the fraction of the condensate which causes particle
growth depends on several factors and ranged from about 0.1 to 0.4
of the total condensate for the range of parameters we explored.
     If one is concerned only with diffusiophoretic deposition
the particle collection efficiency would, therefore, be 60% to
90% of that computed without accounting for condensation on
particles.  In the case of a scrubber which also employs inertial
impaction the particles would be agglomerated to some extent by
the diffusional sweep, so they would have higher mass and be
easier to collect.
     Without going into a detailed model of this phenomenon one
could use either of two simplifying assumptions :
     1.  Assume that the condensation on particles causes no
         agglomeration.
     2.  Assume that the condensation on particles causes
         agglomeration and that the inertial impaction
         efficiency is sufficiently  high that all of the
         particles swept to other particles will be collected
         by impaction.
The first assumption will lead to too low an efficiency and the
second to too high an efficiency.  However, the maximum differ-
ence between the two for a representative case of 25% of the
volume condensing and 25% of that going to the particles would
                             146

-------
be  5.3%  (i.e.,  0.25  x  0.85  x  0025  x  100).   This  is  a  relatively
small  effect  compared  to  the  other uncertainties.
PARTICLE GROWTH
     Particle growth is dependent  on how well  the particles  can
compete  with  the  cold  surface for  the  condensing water.   There
are  several transport  processes  at work simultaneously  in the
condenser  section of an F/C scrubber:
      1.   Heat transfer
          a.   From the  gas to  the cold  surface
          b.   From the  particles  to the gas
      2.   Mass transfer
          a.   From the  gas to  the cold  surface
          b.   From the  gas to  the particles
     A mathematical  model which  accounted  for  these transport
processes  in  addition  to  particle  deposition has been described
in  EPA reports  2'3'**.   The  portions  of that model relating to
particle deposition  were  deleted to  provide a  model which would
describe particle growth  in the  absence of  deposition.   The
basic  relationships  involved  are as  follow:
     The rate of  change of  particle  radius  is  given by  a mass
balance,
                             (PG  "  Ppi} .  cm/s                 (3)
where :
                 2  D   P
                          =  particle  to  gas  mass  transfer  coeffi-
          pG    RI.Q  d   p       cient,  gmol/cm2 -s-atm             r.^
                   f                °                          ^4j
       pG   =  water vapor  partial pressure  in  bulk  of  gas  bubble,  atm
       pRM   =  mean partial pressure to non-transferring  gas,  atm
       r     =  particle  radius,  cm
        P     F
       Tr   =  gas  bulk  temperature, °K
             = molar  density  of water,  gmol/cm
                                              3
             =  water  vapor  partial  pressure  at  vapor-liquid  inter-
               face,  atm
                              147

-------
     Particle temperature can be computed from an energy balance:
     V
                                               CPG - Ppi)    (5)
where:
      2k
h r = -j-
 p     p
     =  particle to gas heat transfer coefficient,
       cal/cm2-s- °K
                                                             (6)
     where  C   = heat capacity of particle, cal/g-°K
              k = thermal conductivity of gas, cal/cm2 -s- °K/cm
             L,. = latent heat of vaporization for water, cal/gmol
              t = time, s
     The overall energy balance for the gas-liquid interface is
given by :
  G  at LM (PG '  L    p
                         } A  dZ =
hL at
                     Ap dZ + hG at  (TLi - V Ap dZ
      where
              k'   = mass transfer coefficient,  gas to liquid,
                    gmol/cm2-s-atm
              a   =  interfacial area  for  transfer
                   volume  of scrubber, cm2/cm3
                 =  cross-sectional area  of  scrubber,  cm2
        A
        h
                 =  heat  transfer coefficient,  gas  to  liquid,
                   cal/cm2-s-°K
              I,  =  temperature  uf liquid bulk,  °K

      The equations given above are used along with enthalpy and
 material balances  for the total system of gas, liquid, and sus-
 pended particles to form a mathematical model for condensation
 and growth.  The model was solved through a finite difference
 method on an electronic computer for several situations which
 are discussed below.
                               148

-------
PREDICTION OF CONDENSATION
     The condensation model was used to predict the particle
condensation ratio, f , (which is the fraction of the total
condensate which goes to the particles) as a function of several
parameters.  The conditions investigated are as follow:
     1.  Scrubber type - one sieve plate
     2.  Inlet gas - saturated from 310°K to 350°K
     3.  Water - uniform bulk temperature from 310°K to 325°K
     4C  Particle number concentration - 107 to 109/cm3
     5.  Particle diameter - 0.1 to 1.0 ym
     6.  Liquid phase heat transfer coefficient - 0.01 to 0.1 cal/cm2-s-(
     7.  Condensation can occur when the gas is saturated
     The computed values are plotted on Figures 3 through 7.  As
can be seen, the figures show the following:
     (Figure)  3.  "^D" does not depend much on "d "
               4.  "f " decreases significantly with "TT"
                     P                                 L
               5.  "f " decreases with "T " to an extent which
                     P                   u
                   depends on "T,"
               6.  "f " increases slightly with "n " above 107/cm3
               7.  "f " increases with "hL" up to "h, "
                   ~ 0.1  cal/s-cm2-°K
     It was  found in computations not reported here that "f "
decreases  significantly with "n  " below about  106 particles/cm3.
Industrial emissions generally have particle number concentrations
on the order of  107, and greater.  The liquid  phase heat transfer
coefficient  is an important parameter but unfortunately, predic-
tions of its magnitude vary considerably depending on which
correlation  is used.  The value  of 0.1 appears to be the best
supported  by the literature  for mass transfer on perforated plates.
     For a combination of parameters such as might be encountered
in a practical situation a value of "f " - 0.25 appears to be
reasonable.  Given this one can  compute the amount of particle
growth that will result from a given condensation ratio
(i.e., g water condensed/g dry gas = q" = condensation ratio).
If the particle size distribution and the scrubber character-
istics are known one can predict the overall penetration that
will be achieved in the scrubber.
                             149

-------
INERTIAL IMPACTION DURING BUBBLE FORMATION
     During the formation of bubbles on a sieve plate the jets
of gas emerging from the perforations impact on the liquid.
Particles are thus deposited on the liquid surface by inertial
impaction.  Particle collection can be determined from:
                           40 F2d2 p  C' u,
              Pt. =
     where    F = foam density, volume fraction liquid
             d  = particle diameter, cm
             p  = particle density, g/cm3
             C1 = Cunningham slip correction factor, dimensionless
             u,  = gas velocity in the perforation, cm/s
             Up = gas viscosity, poise
             d,  = diameter of perforation, cm
            Pt • = penetration of particles of diameter d ,  fraction
PERFORMANCE PREDICTION METHOD
     The sequence of steps to be followed in predicting the perfor-
mance of an F/C  scrubber system involving a sieve plate column
is as follow:
     1.  Determine the initial particle size distribution.
     2.  Compute particle penetration from the saturator based
on the saturator collection efficiency characteristics and the
initial particle size distribution.  No growth occurs in the
saturator .
     3.  Compute particle penetration due to inertial impaction
during bubble formation on the first plate.  Use  the
particle size distribution leaving the saturator  and the collec-
tion efficiency relationship for sieve plate given in equation
(8).
     4.  Calculate the condensation ratio corresponding to the
scrubber operating conditions, from this compute  "f ", the
volume fraction of gas condensing, and then calculate the
penetration due to dif fusiophoresis according to  equation  (9)
                              150

-------
 for  a  conservative  estimate  or  equation  (10)  for  an  optimistic
 estimate
                    1 -PtD =  0.85  (fy)  (l-fp)                  (9)

                    1- PtD =  0.85  £y                           (10)
     The diffusiophoretic penetration applies equally to all
particle sizes so it will not change the size distribution but
will decrease the particle concentration.
     5,  Determine the particle size distribution leaving the
condenser from the values of "q1" and "f ".  Figure 8 is a size
distribution plot showing lines for the particles before and
after condensation.  The conditions used for this plot were:
     Initial d   =0.75 ymA = dry mass median diameter
              r o
              a  = 2.5 = geometric std. deviation
              n  = 109/DNcm3 (Dry Normal cm3, @ 0°C)
              q'  = 0.3 g/g
              fp = 0.25

     6.  Compute the particle penetration function for the
remaining stages of the scrubber, based on the penetration for
one stage given by equation (8).   The penetration for a given
particle diameter on one stage is "Pt-". For "N" stages of
equal efficiency the penetration for a given particle diameter
is "Pt.N".
      1                                   N"
     7.  Use the relationship between "Pt-    and "d " from step 5
and the grown size distribution from step 5 to compute the
overall penetration due to inertial impaction after growth.
     8.  To summarize,  the total  overall penetration for the
F/C scrubber Ft will be the product of the  following:
     a.  "FT " due to impaction in the saturator
         	a
     b.  "Pth" due to impaction in the condenser
     c.  "Ftf " due to diffusiouhoresis in the condenser
           _c
     d.  "PtV' due to impaction in stages after the condenser.
     Thus,
     Pt = PTa x Ptb x PTc x PTd                               (11)

                             151

-------
COMPARISON WITH DATA
     Prediction? based on the model described above were made
for a number of experimental determinations on a 100 m3/min
sieve plate F/C scrubber.  Some examples of comparisons of
predictions with experimental results are shown in Figures 9
and 10,  The correlation shown in Figure 9 is the best obtained
and that in Figure 10 is representative of the average case.
CONCLUSIONS
     The revised model is much simpler to use than our previous
version and it appears to give very good predictions.   It also
offers the opportunity for easy modification to suit specific
situations.  The fraction condensing on the particles  is a key
quantity and it needs more study for plates and other  contacting
systems.

REFERENCES

1.  Calvert, S., J. Goldshmid, D. Leith, and D. Mehta.
    Scrubber Handbook.  A.P.T., Inc., EPA Contract No.
    CPA-70-95.  NTIS No. PB 213-016.  August, 1972.
2.  Calvert, S., J. Goldshmid, D. Leith, and N. Jhaveri.
    Feasibility of Flux Force/Condensation Scrubbing for
    Fine Particulate Collection.  A.P.T., Inc., EPA
    Contract No. 68-02-0256, NTIS No. PB 227-307.
    October, 1973.
3.  Calvert, S., and N. Jhaveri, and T. Husking.  Study of
    Flux Force/Condensation Scrubbing of Fine Particles.
    A.P.T., Inc., EPA Contract No. 68-02-1082. August, 1975.
4.  Calvert, S., and S. Yung.  Study of Horizontal Spray
    Flux Force/Condensation Scrubber.  A.P.T., Inc., EPA
    Contract No. 68-02-1328, Task No. 10.  July, 1976.
5.  Whitmore, P.J.  Diffusiophoretic Under Turbulent
    Conditions, Ph.D. Thesis, University of British
    Columbia, 1976.
                             152

-------
              WATER
              WATER
              WATER
                          CLEAN GAS
                       55°C
                       0.12g/g
                          IMPACTOR
                         0 SAT .A GAS
                       55 C   ^
                       0.12g/g
                           CONDENSER
                           SAT .A GAS
                       74°C
                       0.36g/g
                           SATURATOR
                           HOT A GAS
                       1,000°C
                       O.Olg/g
T
Figure  1.  Generalized F/C  Scrubber System,
                    153

-------
                            CLEAN GAS
            COLD
            WATER
(Imp.)


 Imp.)

(F/C § G)
 (Imp.)
       WATER
                                WATER
Figure  2.   Multiple  plate F/C  scrubber  system.
                      154

-------
•P
U
rt
    1.0
             1     1     1     1     1    1 ----- 1 - 1 - 1
              P
                  345°K (72°C)

                  109/DN cm3

                  0.1 cal/s-cm2-°K
a;

2
o
I—I
C—I
<
00
C
u
    0.5
                                    = 310°K
E-
                                    = 320°F (47°C,
    0.1
             1     1     1     1     1     1     1    1     1     1
        Figure  3.
                                0.5                      I . 0  1

                        PARTICLE DIAMETER,  ym


                   Effect  of particle diameter on condensation

                   rat io .
                                                                . 1
                             155

-------
c
o
u
OS
OS


2
O
HH

H
2
W
Q
W

u
h—I
H
os

On
     1.0
    0.5
    0.1
             n   =
—    P
    d   =
 I         I

109/DN cm3

345°K

0.5 ym

0.1 cal/s-cm2-°K
                   1
                    1
                     1
       305
    Figure 4.
         310      315        320

             LIQUID BULK TEMPERATURE,
                                                320

                                               °K
                                         330
      Effect of liquid bulk temperature on condensation
      ratio.
                           156

-------
    1.0
o
•H
•P
U
03
O

H
2
O
CO

w
o

o
o

w
_}
U
    0.1
             n  =
              P
             d  =
                  1          I

                  109/CN cm3

                  0.5 ym

                  0.1 cal/s-cm2-°K
                                     = 310°K
                                = 320°K
                  1
                            1
1
1
        330
   Figure  5.
                 335         340       345      350

                     GAS INLET TEMPERATURE,  °K
                   355
               Effect  of gas  inlet temperature on condensation
               ratio.
                            157

-------
    1.0
c
o
• H
•P
U
rt

4-1
O
^
H

OS
CO
z
w

z
o
U

w
K-l
o
I—I
H
OS
    0.5
    0.1
             d  =
        —    P
                  I  I | I I II I

                  0. 5  ym

                =  345°K

                =  0.1  cal/s-cm2-°K
                                      1  1  1 1  II 1 1
                                     = 310°K
                                     = 320°K
              1    1  1 1  1 1 1 ll
                                      1  1 1 l I I l 1
        107                  108                 109

             PARTICLE NUMBER CONCENTRATION, #/DNcm3


   Figure 6.   Effect of particle number concentration on
              condensation ratio.
                         158

-------
•P
U
03
Qi

2
O
t—i
H
<
CO
z
w
Q

O
U
    0.1
0.5
   0.1
                  T
                   T
         n
0.5 ym

345°K

109/DNcm3
             I
              I
     I
I
I	I
                                        320°K
I
I
       0    0.01                 0.05                      0.1

               LIQUID  HEAT  TRANSFER COEFFICIENT, hT
                                                  \-i


    Figure  7.  Effect of liquid heat transfer coefficient
               on  condensation ratio.
                         159

-------
3.0
                        I    I   I   I   I    T
1.0
            GROWN
UJ

H-1
Q

tu

cj
I—I


OH
0.5
                               INITIAL
                                          PS
                                          np
                                          q1
                                  0.75
                                  109/DNcm3
                                  0.3
                                  0. 25
0.1
           I
                              1   I   I   I   I
I
I
  Figure 8
                 10        30      50      70

                       DRY MASS  %  UNDERSIZE
                                      90   95
Particle size distribution  before  and after
condensation.
                                                              98
                               160

-------
     0.5
o
• H
*->
O
rt
o
I—I
H
W
2;
P4
P->

W
H-]
O
i—i
H
Pi
             1    I   I  I  I
                               I     I
     0.3
     0.2
0.1
    0.05
                    EXPERIMENTAL

                       PREDICTED
    0.03
              I   I   I  I  I I I
   0.3      0.5
                            1.0
2.0    3.0
                  PARTICLE DIAMETER, ymA
   Figure  9.   Particle  penetration versus aerodynamic
               diameter  for Run 64.
                        161

-------
     0.5
 o
 rt
o
t-H

H



H




0,




u
I—I

H
c£
     0. 3
     0.2
     0.1
    0.05
    0.02
        0.3
             T    I   I   i  I  I
                         EXPERIMENTAL


                            PREDICTED
             I    I   I  I  I  I
               0.,
1.0
2.0    3.0
               PARTICLE DIAMETER, ymA
Figure 10.
           Particle penetration  versus  aerodynamic
           diameter for  Run  69.
                     162

-------
                INTERFACIAL SURFACE EFFECTS ON PARTICLE COLLECTION
                                  G. Woffinden
                                    D. Ensor
                                  G. Markowski

                           Meteorology Research, Inc.
                              464 W. Woodbury Road
                           Altadena, California 91001
                                    L. Sparks

                   Industrial Environmental Research Laboratory
                         Environmental Protection Agency
                   Research Triangle Park, North Carolina 27711
                                     J. Berg

                       Department of Chemical Engineering
                            University of Washington
                            Seattle, Washington 98105
                                    ABSTRACT

     A theoretical analysis was made to show the effects of interfacial sufrace
tension on collection efficiency of solid participates by liquid water droplets.
For wet scrubber applications, it was assumed that particulates had diameters
ranging from 0.05 to 3 urn, and collecting droplets had diameters ranging from
20 to 200 urn.
     The theory indicates that as a particle and droplet impact, the air or vapor
layer between them takes time to flow out.  If the film thinning time is less
than the impact time (particle stopping time), the particle and droplet will
coalesce.  If the thinning time is longer, the particle will rebound and will not
be collected by the droplet.
     Initial theoretical results indicate that, after impact, capture (coales-
cence) will occur in most cases that are considered to be representative of a
typical wet scrubber.  The primary effect of reducing droplet surface tension
is a deeper penetration by the particle into the droplet before stopping because
of the lower resistive force.  The deeper penetration increases film thinning
time and, therefore, may actually reduce collection efficiency.
     Experiments are being conducted to evaluate the results of the analytical
study.  Two types of experiments are being considered: (1)  ultrahigh speed photo-
graphy of the coalescence process between individual particles and droplets, and
(2) evaluation of surface tension variation effects in a laboratory-scale simu-
lated scrubber.
     Initial theoretical work has been completed.   Experimental work is currently
being planned and conducted.  A final analytical study will relate results of
the theoretical mechanism study and the experimental evaluations to current
scrubber modeling techniques.

                                       163

-------
    INTERFACIAL  SURFACE  EFFECTS ON PARTICLE COLLECTION
The objective of this study was to evaluate the effect of interfacial surface
tension on collection efficiency of a wet scrubber.

SCRUBBER MODEL

The effect of surface tension is not obvious in most of the current wet scrub-
ber analytical models1.  Penetration (one minus the collection efficiency) in
the Calvert model,  for example, is given by Equation 1:
                P = exp
                            55 Q
                     (k,f)
(1)
where
         V
         d
         F(k,f)
average  penetration
volumetric flow rate  of gas phase
volumetric flow rate  of liquid phase
velocity of gas phase
viscosity of gas
density of liquid
diameter of liquid drops
function of inertial parameter,  k,  and unknown
factors,  f
Surface tension,  as such,  is not one of the parameters listed; however,  sur-
face tension influences droplet breakup and, therefore,  the resultant droplet
diameter,  d.  Surface tension also affects the Stokes number indirectly be-
cause it is a function of droplet size.  The interfacial surface properties can
also have a significant effect on the "F"  factor because it is primarily a col-
lection efficiency factor that includes both collision and coalescence proba-
bilities.  It is assumed that surface tension does not have a direct effect on
collision probability, but that  it could have an effect on coalescence after
impact.

Limited experimental studies  on the effects of surfactants in scrubbers have
been performed by Bughdadi  .  He concluded that addition of surfactant
(0. 1 percent Triton CF-10) improved the collection efficiency of a venturi
                                   164

-------
scrubber,  especially at low liquid-to-gas flow ratios.  He reported that
the overall collection efficiency improved from 99. 66 percent to 99. 93
percent at 1&1 water/28 m3 of gas (4 gal/1000 ft3).  He attributed the im-
provement to easier penetration of the collected particulates  into the
scrubber water droplets, thus providing more effective wetting.  He also
observed the reduction in spray droplet sizes with surfactant additives.

Wolfe3 has shown experimentally that spray droplet size is  proportional to
the square root of the surface tension:

                        d  = ccr1/2                             (2)

where       d   =   droplet diameter
             c   =   constant
             °   =   liquid surface tension

The maximum surface tension of scrubber water is approximately 72 dynes/
cm (pure water).  Additives and normal contaminants in operating scrubbers
will typically reduce the surface tension to 50-60  dynes/cm.  If a surfactant
is added, the surface tension can be further  reduced to 20-30 dynes/cm.
Therefore, the maximum possible  reduction, from 72 to 20 dynes/cm, will
reduce the spray droplet diameter by only 50 percent at most.  In actual
practice, a maximum reduction of 30 percent is more realistic.  The effect
of droplet  size reduction on scrubber efficiency is not easily  determined from
the scrubber model, Equation 1.   The droplet size affects the "F" factor
which includes individual droplet collection efficiency and other unknown
factors.  Decreasing the size of droplets will increase the number of drop-
lets so that droplet/particle  collision efficiency will be improved,  particularly
for small particles.  When the  droplets get small enough that they are no
longer removed from the gas stream by the entrainment separator, the scrub-
ber efficiency will become  very low.  The droplets in a scrubber are usually
very large (on the order of 50 to 200  |a.m diameter) compared to the particles
being scrubbed (the most difficult particles to remove are typically 0. 05 to
3. 0 jj,m diameter).  When there is a large difference in size between the
droplet and particle and the particle is in close proximity to the liquid sur-
face,  the droplet surface can be considered as an infinite  flat plane.  In
general,  a reduction in droplet size tends to reduce penetration of particles,
i.e.,  it increases the scrubber collection efficiency.  The changes, however,
are expected to produce second order effects.

THINNING LAYER COALESCENCE MODEL

As already indicated, particle capture by liquid droplets is a two step pro-
cess:  (1) collision,  and (2) coalescence after collision. There is virtually

                                  165

-------
no effect of surface tension on collision probability because surface tension,
by itself,  produces no forces that extend far beyond the surface.  The  surface
tension of the droplet, however, can have significant effects on the particle
dynamics at close range, i.e., at the time of collision.

Analytical models relating the physical and chemical mechanisms that con-
trol coalescence were reviewed and compared.  The film thinning model is
the most widely accepted theory of coalescence.

The film thinning model assumes that as a particle collides with a water drop-
let,  a thin layer  of air or vapor is trapped between the particle and the de-
formed surface of the droplet.  This  air layer prevents immediate coalescence.
The layer thins under the forces  involved in  the  collision processes;
thinning is opposed by the layer's viscous forces.  If the layer thins  suffi-
ciently before the colliding  particle rebounds, the layer  ruptures and coa-
lescence occurs.  Figure 1 illustrates the collision and film thinning model.
The solid particle has penetrated its maximum distance, xm, into the  sur-
face of the much larger water droplet, and an air layer has been trapped
between the particle and the droplet.  Fs denotes the force between the
particle and droplet and depends on a number  of parameters, including
the droplet surface tension,  internal pressure, particle  velocity, and particle
radius.  The particle/droplet contact time during the collision (before re-
bound) is calculated either analytically or numerically, depending on the
assumptions made.   The contact time is then  compared with  the thinning
time.  If the thinning time is less than the contact time,  coalescence is
assumed to occur.

The time for the trapped layer to thin is calculated using Reynold's formula":
                    t=  -1 A*.  „  /J__A\                 (3)
                        ^ FS    U   *J
where       t    =    thinning time
             A   =    area of contact
             ij    -    bulk viscosity of air
             b    -    film thickness at maximum particle penetration
                      distance
             ^0   =    initial film thickness
             F   =    force between the particle and the droplet
                                  166

-------
If 6Q  is larger than 25,  its effect on the thinning time,  t, is small.  It is,
therefore, normally neglected.

Results of the current study indicate that, although general trends in coa-
lescence probability may be predicted by the model, modifications will be
required for specific quantitative predictions.  The following model  restric-
tions could account for the observed differences:
1.    Calculations performed to date (Emory/  Lang,6 Jayaratne7) have
used the average force acting on the particle during collision and the area at
maximum penetration in Equation 3 to calculate thinning time, t.  However,
the thinning time is  actually dependent on the time average ratio of the square
of the contact area,  A2 , and the force, Fs,  exerted by the liquid  surface.
Models examined calculated the force as:


                       F  = aAa                             (4)
                        s

where      p     =    force  exerted by the liquid surface
            a     =    a constant depending on velocity
           a         1 or 1/2
           A     -    contact area

Thus,  the  thinning time is estimated as
where CQ = constant, including all other constants in Equations 3 and 4.  Before
the particle forms a trapped air layer, the area,  A, is  zero and the thinning time
also becomes zero since a  is  1 or less.  Thus,  the model appears to  predict
that the thinning layer collapses before  penetration begins.  Of course, the
layer does not collapse completely, but the above argument implies that the
layer thins much more rapidly  at the start of penetration than  is accounted for
by the ratio of average force and  maximum contact area.
2.    The calculated thinning time is strongly dependent on the assumed shape
of the depression created by the impacting particle.  Figure 2 shows two models
that have been used.  The area of contact is proportional to  62  and the thinning
time is  proportional to 04.  The difference in thinning time between these models
is likely to be an order of magnitude.   Model 2 also predicts a smaller droplet
resistive force and, thus,  a considerably longer contact time before rebound.
                                  167

-------
3.    Thinning time is strongly dependent on  6, the separation at which sur-
face forces cause rupture of the film.  Uncertainty exists in estimating this
                                                                  o
unstable film thickness.   Jayaratne  suggests a thickness of  1000  A,  and
^^                                            o
Emory5  indicates that it may be as thin  as  50 A .  This factor  of 400 differ-
ence introduces uncertainty in thinning time.

4.    The noncontinuous nature of the air film, when its thickness is less than
the mean free path of approximately 1000 A is expected to decrease its stabil-
ity, and  therefore decrease the thinning  time.  Any electrostatic charge or in-
duced molecular polarization is is also expected to reduce film stability.

5.    Any surface roughness in the colliding particle can shorten the thinning
time.  Analysis of Equation 3 indicates that the thinning time is sensitive to
roughness  either through reducing the effective area,  A, or the rupture dis-
tance,  5 .

6.    Collision at a  small grazing angle  causes a particle to deflect rather
than rebound in the opposite direction.  Many models have not attempted to
include this effect.  It is clear that the impact angle is important; however,
the overall effect in actual scrubber applications is not obvious.

7.    Davis8  predicts a marked decrease in film viscosity when the thick-
ness is less than the mean free path.  In this case,  the effective viscosity is
nearly directly proportional to the film thickness.  When this  effect is in-
cluded in the calculation of thinning time from Equation 3,  the thinning time
is reduced by a factor on the order of 5.

8.    The electric double  layer at a water surface produces the effect  of a
small negative surface charge.  Therefore,  a short-range repulsive force,
proportional to the dipole  moment, is generated between two droplets
in close  proximity9.  Silica particles adsorb a water layer from the atmos-
phere and, therefore, acquire an electric double layer similar to that  of
a water droplet.  For initial stages of coalescence, they, therefore, can
be considered as water droplets.  Double-layer forces generally are estimated
to be smaller than those produced by the viscous film thinning effects10.  There
is evidence,  however, that the double layer may have a. coalescence retarding
effect under some conditions.  This effect was used to explain  stability of
water droplets floating for long periods on a water surface'1 .  Similar  experi-
ments  have been conducted in our laboratory.  Droplets of polar liquids are
generally,  but not always,  more stable floaters than those of nonpolar liquids.
Other mechanisms,  such as electroviscous effects or gas adsorption on the liquic
surface, may have an effect also.  It has been shown that addition of surfactants
                                  163

-------
can increase the stability of floating droplets. This effect could be the
result of increased polarization of the surface double layer.  The effects
from surface  polarization must be treated separately from those due to
surface tension alone.

If the film thinning model or a modified thinning model is valid, it predicts
that the surface tension of scrubber droplets will have only a secondary
effect on particle collection efficiency due primarily to reduced size of
scavenger droplets.  A reduction in surface tension reduces  the pressure
exerted on the particle during impact. Since the particle will then have
less resistance, it penetrates deeper into the droplet. This deeper pene-
tration  produces a longer,  narrower path for the vapor film, and the thinning
time  is increased. The increased thinning time results in reduced capture
probability because the particle may rebound before the film ruptures.

EXPERIMENTS

Simple laboratory experiments were devised to measure film thinning times
and the effects of surface energy  on the thinning times.  These experiments
were planned  to evaluate, at least qualitatively,  the film thinning theory.

Initial experiments have been performed, and additional experiments are in
progress.  In these experiments,  the impact and coalescence process for
a water droplet and a glass sphere (representing a fly ash particle) was ob-
served with a high speed motion picure  camera looking through a microscope.
The water droplets were approximately 1000 to 3000  (o.m diameter and
were suspended on the end of a microliter syringe needle. The glass spheres,
were made by drawing a glass rod into a fine fiber and forming a ball on the
tip end.  The  particles ranged from 40 to 4000  |am diameter. The  suspending
rods were as  small as 10 jj.m diameter.  Rods 10 jj.rn diameter without a
sphere on the end were also used  to represent 10 p.m diameter particles.
The experimental setup is shown in Figures 3 and 4.  The supporting syringe
and glass rod were mounted on three-dimensional micromanipulators  so that
they could be  moved independently within the field-of-view of the camera.
The glass particle was caused to impact the water droplet by rapidly advanc-
ing the horizontal traverse mechanism of the particle support,  Figure 5.
Radioactive polonium strips were mounted near the particles and droplets to
eliminate electrostatic charges.  Radiation from the polonium ionizes the  air
so that surface charges bleed off.  Experiments without the polonium tended
to give shorter and more variable film thinning time measurements. The
camera used was  a Beckman and Whitley Dynafax, operating at  a top speed
                                  169

-------
of 26, 000 pictures per second, with individual frame exposure times of 2. 5
jisec.  The camera  is a continuous writing, rotating drum camera with 224
frames.

Illumination was provided by a Xenon flash lamp system.  The system is
designed to provide  a square-wave light pulse with nearly uniform intensity
over the picture taking time of 1 drum revolution.  In order to eliminate
the effects of heat and electromagnetic fields produced by the flash lamp
discharge,  a fiber optic light pipe was used.   The light pipe made it pos-
sible to move the lamp away from the event being  photographed.

Coalescence delay time was measured as a function of particle size.   The
effects of impact velocity and surface tension are  to be determined in future
studies.   Figures 6  through 12 are typical photographic sequences showing
the coalescence process.  Film thinning time was taken as the coalescence
delay time; i. e. , the time from first contact until a liquid meniscus  was
first discernable.  For example, in Figure 10 the delay time was taken as
12 frames at 39  ^sec/frame, for a total delay time of 468 (j.sec.  The
39 (J. sec is  determined by the framing rate, which was normally held con-
stant at 26, 000  frames/sec, except for Figure 6 which was taken at 10, 000
frames/sec, giving  a time per frame of 100  ^sec.  The magnification in
Figure 6 is also reduced.

The camera speed was measured to within ±5 percent.  A measurement
uncertainty of ±1/2  frame at each end of the thinning time measurements
introduces an additional potential error of  ±39  jisec.  A thinning time of
1, 000 p. sec, therefore,  has a random experimental variation of  approxi-
mately ±10 percent  while a thinning time of 100  JJL sec has a random vari-
ation of approximately ±50 percent.

Figure 1 3 is a comparison of experimental results with theoretical pre-
dictions based on the film thinning model of Emory5 .  The experimental
results are consistent and reasonably reproducible; however, they differ
from the theoretical predictions by as much as 2 orders of magnitude
for larger droplets.  For the large particle experiments,  the particles
were nearly the same size as the water droplets.  An error could, there-
fore,  be introduced  by the false assumption that the droplet was  large com-
pared to the solid particle.  Some of the random experimental variation
could also be due to uncontrolled parameters including: ambient temperature,
relative humidity, water purity, and particle cleanliness.  No special control
of these factors was made; however,  because of the limited ranges of varia-
tions they are not expected to have large effects. Ambient temperatures
ranged from 19  to 22°C  (67-72°F).  Relative humidity ranged from 40 to
                                  170

-------
49 percent. Freshly distilled water was used for the droplets. Glass particles
were initially clean when formed on the end of glass  support rods in a torch
flame. They were allowed to remain in the ambient atmosphere for at least
several days with no additional cleaning or treatment prior to use in the
experiments. In many cases, the particles were used repeatedly  in successive
tests. Water was blotted from the bead surface at the end of a test run,
and the bead was allowed to dry in the nominal ambient 45 percent relative
humidity atmosphere for at least 15 minutes prior  to performing  a subsequent
test  with the same particle. An adsorbed monolayer  of water probably existed
on the particle surface for  all tests, since it is almost impossible to preclude
such a layer in  an ambient  atmosphere.  Also, a similar water layer probably
will  exist on fly ash particles in a scrubber.

CONCLUSIONS

1.    An impacting  solid particle appears to remain  in contact with a water
droplet for a finite measurable time before wetting or coalescence occurs.
This  coalescence delay time can be interpreted as  equivalent to the film
thinning time.

2.    The  film thinning models of droplet coalescence and experimental re-
sults from the MRI  studies show the same general trends even though there
are large differences under some conditons.  At least some of these dif-
ferences,  hopefully, can be resolved by additional  development of the model
and the experiments.

3.    The  film thinning model is probably the best model available, even
though it appears that modifications are needed to provide more valid pre-
dictions.

4.    Reduction in surface tension of the scrubber liquid appears  to produce
only second order effects on collection efficiency; i. e. , through reductions
of spray  droplet sizes, or through reduced penetration resistance during
particle impact. The magnitude of these effects is not yet known.

5.    Electrostatic fields appear to be effective in increasing coalescence
efficiency.  Charge effects  can produce large reductions in film thinning
times.  Therefore,  addition of electrostatic fields in a wet scrubber could
probably increase the particulate collection efficiency.


ACKNOWLEDGMENTS

Gratitude is expressed to Patrick A. O'Donovan, Aerojet General Corp-
oration,  for providing high  speed photographic equipment.


                                  171

-------
Figure 1.   Model for collision of particle and liquid droplet.
                        172

-------
     Model 1
Model 2
Figure 2.  Models for surface deformation.
                173

-------
                                       I
                                       CO
                                       V
                                       f— I
                                       (0
                                       O
                                       o
                                       fi
                                       o
                                       •43
                                       §
                                      •a
                                       O
                                       o>
                                       4>

                                      |

                                       cr
                                      o
                                      y
                                      CQ
                                      o


                                      is
                                      V  CO
                                      fl  -3
                                      •i-<  G
                                      y  ctt

                                      •s  «
                                      X  «
                                         o
                                      43  O
                                      60  fl
                                      0)
                                      fn
                                      3
                                      W)
174

-------
                                                            77-083
Figure 4.  Glass rod with simulated fly ash particle mounted on tra-
           versing mechanism.
                                  175

-------
                                                         CQ
                                                         +J
                                                         
-------
     XMM
                                             77-115
Figure 6.  Coalescence of 2800 \i m diameter glass particle
           with water droplet,  1200 ^sec delay (10, 000 pps).
                             177

-------
                                             77-116

Figure 7.  Coalescence of 2800 pm diameter glass particle
           with water droplet, 2067 M sec delay.
                            178

-------
                                77-117
Figure 7.  Continued.




               179

-------
     ••^•"""•"••^••••••w     ^^^^^^»^^^««™i^^^»     gm^u^m    MHil^^^
                                                  77-118

Figure 8.  Coalescence of 1700 Jim  diameter glass particle
            with water droplet,  585 ;isec delay.
                                180

-------
                                               77-119

Figure 9.  Coalescence of 1000 /Lim diameter glass  particle
           with water droplet, 780 flsec delay.
                             181

-------
                                             77-120

Figure 10.  Coalescence of 725 Mm diameter glass particle
            with water droplet, 468 jUsec delay.

-------
                                             77-121
Figure 11.  Coalescence of 275 Jim diameter glass particle
            with water droplet, 351 Jisec delay.
                           183

-------
                                           77-122

Figure 12.  Coalescence of 100 pm diameter glass particle
            with water droplet, 156 Jisec delay.

                             184

-------
 10,000r
  1,000  —
CO
e
o
fH
a
• H
V
4->
0)

s
ni

q

0)
I—I
o
     100
        Figure 13.
 Coalescence Delay Time (sec. )



Comparison of theoretical predictions and exper-

imental measurements of coalescence delay time.
                                185

-------
REFERENCES

1.    S.  Calvert, J. Goldshmid, D. Leith and D.  Mehta, "Wet Scrubber
      System Study, Volume I, Scrubber Handbook," APT, Inc.  Report
      No. EPA-R2-72-118a, NTIS No.  PB 213016, 1972.

2.    S.M. Bughdadi,  "Effect of Surfactants on Venturi  Scrubber Particle
      Collection Efficiency," M.S. Thesis, Department of Thermal and
      Environmental Engineering, Southern Illinois University of Carbon-
      dale, 1973.

3.    H.E. Wolfe and  W.H.  Andersen,  "Kinetics,  Mechanism, and Result-
      ant Droplet Sizes of the Aerodynamic Breakup of Liquid Drops,"
      Aerojet-General Corporation Report No.  0395-04(18)SP, 1964.

4.    O.  Reynolds, Phil. Trans. Roy.  Soc. ,  A(l 77):157 (1 886).

5.    S.  Emory and J.  Berg, "The Effect of Liquid Surface Tension on
      Solid Particle-Liquid Droplet Coalescence," University of Washing-
      ton, Task Report EPA Contract No. 68-02-2109.  (To be published)

6.    S.B. Lang, "A Hydrodynamic Mechanism for the  Coalescence of
      Liquid Drops,"  Ph.D. Thesis, Lawrence Radiation  Laboratory
      Contract 7405-eng-48,  1962.

7.    O.  W.  Jayaratne  and B. J. Mason, "The Coalescence and Bouncing
      of  Water Drops at an Air/Water Interface," Proc. Roy. Soc. , A(280):
      545 (1964).

8.    M.H.  Davis,  "Collision of Small  Cloud Droplets:  Gas Kinetics  Effects, "
      J.  Atmos. Sci. ,  (29):911 (1972).

9.    A.  W.  Adamson,  Physical Chemistry of Surfaces,  Second Edition,
      Interscience Publishers, New York, New York, I960.

10.   N.  Arbel and Z.  Levin, "The Coalescence of Water  Drops,  I. A
      Theoretical Model of Approaching Drops," Department of Geophysics
      and Planetary Sciences, Tel Aviv University, Ramat Aviv, Israel,
      February, 1977.

11.   C.L. Strong, "The Amateur  Scientist." Sci. Am.. (230):104 (1973).
                                   186

-------
               EPA MOBILE PARTICULATE COLLECTORS
                   A: PROGRAM INTENTIONS
                        Dale L. Harmon
                 Particulate Technology Branch
         Industrial Environmental Research Laboratory
              Research Triangle Park, N. C. 27711
          B: CUPOLA FOUNDRY PARTICLE CONTROL WITH
               VENTURI AND SIEVE TRAY SCRUBBERS
                         D. L. Zanders
                       S. P. Schliesser
                 Monsanto Research Corporation
              Research Triangle Park, N. C. 27711
                           ABSTRACT
     The following two papers describe the EPA mobile particulate
control units and the results of tests using the mobile scrubber
unit to control the particulate emissions from an iron foundry
cupola.
                              187

-------
                     EPA MOBILE PARTICULATE COLLECTORS
                           PROGRAM INTENTIONS
                             Dale L. Harmon
                      Particulate Technology Branch
              Industrial Environmental Research Laboratory
                   Research Triangle Park, N. C. 27711
                               April 1977
Introduction
     The Environmental Protection Agency and predecessor organizations
have been involved in development of systems for control of emissions
from industrial sources for many years.  The EPA R&D program for particulate
control is designed to establish engineering design techniques and
performance models, and to improve the collection capability and economics
of control devices for particulate matter.  The objective is the develop-
ment and demonstration of control technologies capable of effectively
removing large fractions of the under 3 micron diameter particles from
effluents.

     In 1973, EPA funded the design and construction of a mobile fabric
filter unit, the first of a series of three mobile research units built
for the assessment of the collectibility of particulate emissions from
industrial sources. These units house the three types of conventional
particulate collectors (fabric filter, wet scrubber and electrostatic
precipitator) that are generally used by industry at the present time.

     Actual data on particle collection efficiency of conventional
particulate collectors is sparse.  Actual operating data for the optimiza-
tion of collection efficiency and cost is not readily available.   Design
of control equipment is presently based on projections from historical
data which has been developed by manufacturers for their own devices and
is proprietary.  This information is not standardized and cannot be
extrapolated to other devices.  On-site testing prior to control device
selection is seldom attempted and the possibility of alternative devices
is poorly defined.  The EPA mobile units are highly versatile and will
be used to investigate the applicability of these control methods to the
control of fine particulate emitted from a wide range of sources.  The
relative capabilities and limitations of these control devices will be
evaluated and documented.  This information, supplemented by data from
other EPA particulate programs, will permit selection by equipment users
of collection systems that are technically and economically optimum for
specific applications.  Operation of the mobile units will be coordinated
with other EPA laboratories and regions to provide, when possible, field
data on specific problem sources.  Capabilities of the three mobile
units are tabulated in Table I.

Fabric Filter Unit

     The original mobile fabric filter unit was built by GCA Corporation
for EPA and was mounted on an open 1360 Kg truck.  The equipment could
be operated on the truck or could be removed and operated at locations
not accessible to the truck.  The mobile filter unit can be adapted to
cleaning by mechanical shaking, pulse jet or low-pressure reverse flow.

                                  188

-------
Filtration can be conducted at cloth velocities as high as 0.1 m/s with
a pressure differential up to 50 cm of water, and at gas temperatures
up to 290°C.  Cleaning parameters can be varied easily over broad ranges.
One to seven filter bags of any fabric media, 1.2 to 3m long and up to
30 cm in diameter may be used.  Gas flows used are from 0.012 to 0.13
m/s, as determined by cloth velocity, bag size and bag number.

     The mobile fabric filter unit in the above configuration operated
on effluents from a brass and bronze foundry, a hot mix asphalt plant,
a coal-fired boiler, a lime kiln and a pulp mill recovery boiler.  Operation
of the unit from the open truck was not satisfactory due to problems
resulting from the weather, so the filter system was installed in a 12m
closed trailer in October 1976.  Since then the unit has been used to
determine the performance of a fabric filter on air emissions from a
cyclone collector used on the St. Louis Refuse Processing Plant.  The
unit is scheduled to be operated on a power plant burning a western
subbituminous coal this month (April).

Met Scrubber Unit


     The second mobile unit received by EPA was the wet scrubber.  This
unit, built by the Naval Surface Weapons Center at Dahlgren, Va., was
delivered in December 1974.  The scrubber unit is in a 12m trailer and
contains a venturi scrubber and a sieve tray scrubber.  Maximum gas flow
rate.,through the scrubbers is 0.28 m /s.  Maximum L/G ratio is 0.01
m /m .  The unit can treat gas at temperatures up to 480°C.

     Figure I is a flow schematic of the scrubber unit.  The inlet,
through the side of the trailer, is equipped with four banks of electrical
heaters to prevent heat loss of the effluent stream in the duct.  A
presaturator in the mobile unit precools the gas down below 175°C before
treatment by the scrubbers.  The presaturator can be bypassed if preceding
is not required.  After the presaturator the gas flow goes to either the
sieve tray scrubber or the venturi scrubber.

     The sieve tray scrubber is a four-tray tower designed for ease of
operation and maintenance.  Aluminum support plates secure each tray
holder on three sides.  High-density polyethylene flanges, bolted directly
to the tray holders, are sealed on the holder side with 0-rings.  These
flanges are individually machined to fit 46 cm OD Pyrex glass column
sections, and hold an 0-ring which seals against the outside diameter of
the glass section.  A clearance of approximately 0.16 cm on either end
of each glass section allows the glass to "float" free of compressive
or tensile stresses between tray holders.   Three sets of trays are
provided with varying hole diameters and hole spacing.  The open area
is the same for all trays.  Trays are changed by unscrewing the two
wing nuts on each tray cover, removing the cover and the old tray, inserting
a new tray, and replacing the cover.  One man can change the trays in
about 5 minutes.

     Three interchangeable Venturis with different throat diameters are
provided with the venturi/cyclone scrubber.  The maximum pressure drop
with the smallest venturi is about 250 cm w.c.  Each throat is 30 cm

                                  189

-------
long and accepts water dumped radially 5 cm from the throat entrance.
After leaving the venturi, the scrubbed gas enters a cyclone which
removes large droplets and/or particles and provides a drain for the
scrubbing liquid.

     A demister is located in the duct following the scrubbers to remove
entrained droplets from the scrubbed gas stream before entering the
blower and discharging to the atmosphere through a stack mounted on the
trailer.

     A 1700 liter capacity supply tank provides water for the entire
system. After passing through the scrubbers, the scrubbing liquor drains
to a sump tank mounted under the trailer.  The liquid is pumped from the
sump tank to an Industrial Filters Company deep-bed gravity filter
mounted above the supply tank.

     In both the scrubber and fabric filter mobile unit, a section of
the trailer is partitioned off for a control room and analytical lab-
oratory.

     The mobile scrubber unit has been operated on a coal-fired power
plant, on a lime kiln in a pulp and paper mill and on a gray iron foundry.

Electrostatic Preci pita tor UnU

     The mobile electrostatic precipitator was delivered to EPA in
September 1976. This system was also designed and built by the Naval
Surface Weapons Center.  The mobile ESP facility consists of two separate
units mounted in 12m freight vans.  Figure II is a floor plan of the two
units.  The first unit is the process van which houses a five-section
ESP and all auxiliary equipment.   This equipment consists of a diffuser
section, a blower and cooling system, a screw conveyor for removal of
collected material and transformer-rectifiers for high voltage connections.
Each section of the ESP has a separate power supply which provides up to
50,000 volts and 7 milliamps DC.   The precipitator is designed to operate
at flow rates up to 1.4 m /s at temperatures up to 480°C.

     The second unit is a control/laboratory van containing all process
controls, monitors and recorders plus provisions for an analytical
laboratory and equipment storage.

     The mobile ESP facility is scheduled to be taken to the field for
the first test in May 1977.  It will be used to evaluate the effects of
sodium conditioning on a low-sulfur western coal.

Summary

     The original objective in building the mobile particulate control
units was to operate all three units on the same source to compare the
relative capabilities and limitations of the control methods on a given
source.  In practice the units, more frequently, have been used individually
to answer specific control problems.  These versatile mobile units
should find wide applications in the future to fill both of these needs.

                                  190

-------
                Table I.  MOBILE PARTICULATE COLLECTOR CAPABILITIES
Characteristic

Trailer Length, m

Type Particulate
   Collector
Flow Rate, m /s

Maximum AP,
   cm w.c.
   Mobile
Fabric Filter

     12

1. Shake Clean
2. Pulse Clean
3. Reverse-Flow Clean

0.012 to 0.13

     50
  Mobile
 Scrubber

    12

1.  Venturi
2.  Sieve Tray
0.14 to 0.28°

   250
   Mobile
     ESP

12 (each of two)

1.  5-section ESP



0.47 to 1.4

     66
Maximum Temperature,
°C
Maximum Duct
Length, m
290
30
480
30
480
30
aSieve Tray limited to 0.24 m /s
                                       191

-------
     SCRUBBED
     GAS EXIT
STACK GAS
INLET
                       DEMISTER
       BLOWER
              BLOWER  2
trn
^••Wi
— V
*-|
                                            V///A
                                             SUMP ;
                                            PTANK'
      SUPPLY
       WATER
                                                    PRESATURATOR
                                                        PUMP
        OVERFLOW
                 DRAIN
              M  BALL  VALVE
                  BUTTERFLY
             1X1    VALVE
                  MOTOR  DRIVEN
                  BUTTERFLY
                    VALVE
                      THROTTLE
                       VALVE
                      SOLENOID
                       VALVE
                       CHECK
                       VALVE
                                                   AIR
_    WATER
     PROCESS
     INSTRUMENTATION
     STATION
                      Figure  1.   Scrubber Flow Schematic.

                                     192

-------
                                                        X
                                                        4-1
                                                        o
                                                        03
                                                       a.

                                                       CL.
                                                       LO
                                                       UJ
                                                       4-1
                                                        o
                                                        C
                                                        rt
                                                       r^j

                                                        0)

                                                        p
193

-------
            CUPOLA FOUNDRY PARTICLE CONTROL WITH
              VENTURI AND SIEVE TRAY SCRUBBERS
                             by
              D.L. Zanders and S.P. Schliesser
                Monsanto Research Corporation
This paper shows results obtained from field testing the
Industrial Environmental Research Laboratory mobile pilot
scrubber unit on effluent from an iron foundry cupola.
SCRUBBER SYSTEM

The scrubber unit is mounted in a 40 ft trailer and has a
processing capability of about 500 cfm of effluent gas de-
pending upon operating conditions.  Inside the trailer, two
scrubber units are arranged as shown in Figure 1.   The top
half of the schematic shows the gas flow path - the bottom
half the liquor flow path.

A stack gas slipstream is drawn through the unit beginning
at the top left of the schematic, through a flow measuring
device, presaturator, and into either one of two particle
removal devices.  These are followed by a mist eliminator,
I.D. fans and a scrubbed gas exit stack.

Scrubbing medium is supplied to either unit from the supply
tank.  Liquor returns to the system from the bottom of the
scrubber units into a sump tank, where it is then pumped to
a solids removal filter.  Clarified liquor is returned to the
supply tank.  Provisions are incorporated to treat the liquor
in the filter bed and supply tank areas.  Make-up liquor is
attained through a piping arrangement to the supply tank.

One scrubber is a venturi type, with capability to vary the
throat diameter.  The other scrubber is a four tray tower
with sieve type trays.  This unit has the capability of vary-
ing tray hole size and spacing.  Both units were tested on


                             195

-------
the cupola foundry.

PROGRAM INTENT AND SITE SELECTION

Intent of the testing program was to determine the fine particle
collectability (3 micron and below) of these conventional type
control devices as represented by the pilot scrubber system
and measured by impactors and mass collection devices.

A commonly occuring foundry operation was selected based upon
its previously determined fit within the ranges of over twenty
characteristics of iron foundry operations in the following
areas:

         a.  Feed Metal and Furnace Characteristics

         b.  Charge Characteristics

         c.  Fuel Characteristics

         d.  Effluent Characteristics

The selected cupola was a nominal four ton per hour hot blast
operation producing gray iron castings.

TESTING PROGRAM

Sampling was conducted over a two month period, encompassing
both inlet and outlet particle size distributions and mass
loadings under varying operating conditions for each scrubber
system.

Venturi Unit - Plans

Test plans for the venturi scrubber were to:

         a.  Look at collection performance as a function
             of pressure drop and liquid/gas ratio

         b.  Look at the dependency of collection
             efficiency on throat velocity

         c.  Investigate cooling and humidification
             effects

         d.  Look at the relationship between efficiency
             and pressure drop

A range of operating conditions were covered during the testing
period as shown in Table 1.


                             196

-------
                           TABLE 1

                     Range of Conditions


       Variable

   Pressure Drop (in. H-0)

   Gas Rate (acfm)

   L/G (gpm/kcfm)

   Throat Velocity (ft/sec)


Venturi Unit - Results

Data obtained by varying the L/G ratio at fixed throat velo-
city are shown in Figure 2.  Removal efficiency improved
slightly but steadily through the studied range.

Varying throat velocity affected overall mass collection
efficiency as shown in Figure 3.  The influence here is
considerably more noticeable.

Pressure drop influence is depicted in Figure 4, over the
range of 20-60 inches of water.

The effect of additional humidification was looked at by
operating at the same conditions with and without the use
of the presaturator unit.   This  effect is illustrated in
Figure 5.

A considerable body of data was  gathered on the fractional
removal efficiency of the venturi unit during the course of
the testing program.   This data  includes fractional removal
efficiency curves,  cumulative size distributions and outlet
differential size distributions.  An illustration of commonly
occuring data patterns is shown  in Figure 6.  This graph shows
removal efficiency as a function of particle size from approxi
mately 0.5 to 4.0 microns, as well as the measured effect on
low and submicron particle removal by varying pressure drop in
the venturi.  The wide range of  removal efficiencies measured
indicate a high need to optimize operating conditions for a
venturi type scrubber to attain  the fine particle removal ef-
ficiency potential  of such a device.

Sieve Tray Unit - Plans

The other unit in the EPA mobile scrubber system, the sieve

                             197

-------
tray scrubber, is not commonly applied to control foundry
emissions.  Test conditions covered were similar to the
venturi regarding L/G ratios, but the pressure drop range
was appreciably lower.  Conditions covered for the sieve
tray scrubber are shown in Table 2.
                           TABLE 2

                Sieve Tray Testing Conditions


            Variable                      Range

         Pressure Drop (in. H20)* **      12-15

         L/G Ratio (gpm/kcfm)* **         10-30

              * at two sieve hole sizes

             ** with and without humidification
Overall mass removal efficiency varied from 45-701 by weight
(below the performance of the venturi unit) dependent pri-
marily upon the liquid to gas ratio, with smaller effects
observed from sieve plate hole size and pressure drop across
the column.

Data gathered on the effect of L/G ratio for both small and
large hole sieve trays used (.125 and .250 inch holes) indi-
cated a divergent difference in performance characteristics
as shown in Figure 7.

Examination of pressure drop influence showed fairly large
data scatter cycling between 45-65% removal efficiency.

Humidification with the presaturator produced a slight increase
in the overall collection efficiency of the tray scrubber.

Fractional removal efficiency for the tray scrubber was char-
acterized by a very sharp drop in removal efficiency below
the two micron level, with poor removal efficiency character-
istics on submicron particles.  An illustration of this is
shown in Figure 8.

CONCLUSIONS

The following information and indications have been drawn from
the test data gathered:

                             198

-------
Venturi Scrubber

     a.  Measured overall mass removal efficiency varied
         from 45-96 weight! depending upon operating
         conditions

     b.  Throat velocity significantly affected overall
         mass removal efficiency

     c.  Gas presaturation appears to be beneficial to
         fractional removal efficiency in the low and
         submicron particle sizes, particularly in the
         submicron range

     d.  Measured fractional removal efficiencies as
         high as 70-80 weight! were achieved in the low
         and submicron particle sizes through optimiza-
         tion of scrubber system operating conditions
Sieve Tray Scrubber

     a.   Measured overall mass  removal  efficiency  varied
         from 30-70 weight!  depending on the operating
         conditions

     b.   Fractional removal  efficiency  showed a sharp
         decline below the 2 micron size,  with poor
         removal in the submicron range

     c.   Presaturation of the incoming  gas stream  had a
         slight  beneficial effect on removal efficiency

     d.   Over the range of variables tested, no optimi-
         zation  of fine particle  removal efficiency  was
         attained.
                             199

-------
                       TMWOTTLE VALVE
     TTERFLV VALVE
MOTOR DRIVEN
BUTTERFLY VALVE
                   P>4  SOLENOID VALV
                       CHECK VALVE
                                      I  I PHOCESS INSTRUMENTATION STATIC
Figure  1.    Scrubber  Flow  Schematic.
                             200

-------
  90
   80
H
   70
W
c
s
[Ij

05  60
   50
                         I
         I
I
                10
15      20      25


L/G, gpm/kcfm
       30
    Figure  2.   Effect of L/G ratio or. Venturi Scrubber,



                         201

-------
  100
   90
U
2
W
t— i
u
U-.
PJ
O

UJ
04

O
   80
   70
   60
                     I    T
         I   l   l    I   I
I	I
            250     300     350      400

               THROAT VELOCITY, fps




Figure 3. 'Effect of throat velocity on  Venturi  Scrubber.
                         202

-------
   100
u

§   90
UJ
o
2:
    80
    70
o
    60
               i	r
          i	r
                          I
I
I
                         30        40         50


                         PRESSURE DROP,  in.,  w.c
                    60
       Figure 4.   Effect of pressure drop on Venturi Scrubber




                                 203

-------
 u
 2
 W
 W
 O
 S
 w
   .9
    «
   • o
    6
   -°
 H
 U
 <
 OS
 UH
                       O—NO PRESATURATION

                       A—WITH PRESATURATION
                     I
                    I
.3      .5         1.0

      PARTICLE SIZE,  microns
                                           2.0
Figure 5.   Effect of presaturation on Venturi Scrubber,
                      204

-------
   90
ON=  80

H
u
2
   70
t-L,
w
o
IS
w
DS
60
   50
   40
   30
                                   A—  20" w.c


                                   O—  40" w.c


                                   D—  60" w.c
                          1
                                  I
1
                          234

                       PARTICLE  SIZE,  microns
Figure 6.  Fractional  removal  efficiency for Venturi Scrubber,


                            205

-------
   65
   60
PJ
u.
P-.
UJ
o
Oi
Di
W

O
   55
   50
   45
                        125  in.  sieve
                        250  in.  sieve.
                    I
           I
I
          10
15        20         25

 L/G  RATIO,  gpm/kcfm
          30
   Figure 7.  Effect of  L/G  ratio  on Sieve Tray Scrubber,
                          206

-------
u
I— I
PH
P-,
W
o
s
w
Pi
H
U
<
Pi
PU
    70
   60
   50
   40
   30
                          I
I
                1          2          3

                 PARTICLE SIZE, microns
  Figure 8.   Fractional efficiency of  Sieve  Tray Scrubber,
                            207

-------
                FIELD TEST OF A VENTURI SCRUBBER
                          IN RUSSIA
                     Dennis  C.  Drehmel
               Particulate Technology Branch
        Industrial  Environmental  Research Laboratory
            U.S.  Environmental  Protection Agency
             Research Triangle  Park,  N.C. 27711
     This paper discusses the performance test of a high
pressure drop venturi scrubber operating on an electric
arc furnace at the Nikopol Ferroalloy Plant in the Ukrainian
Soviet Socialist Republic.  The overall scrubber collection
efficiency averaged about 99.9 percent, with greater than
98.6 percent for submicron particles.  These tests were
performed as part of a technology exchange program between
the U.S. and the U.S.S.R.
                             209

-------
              FIELD TEST OF A VENTURI SCRUBBER IN RUSSIA
Introduction

     Because of the adverse effect of pollution on public health and
welfare, the United States of America and the Union of Soviet Socialist
Republics have independently developed pollution control methods to
protect the environment from liquid, solid, and gaseous contaminants. In
1972, with recognition of the mutual benefits that could be gained from
technology exchange, the United States and the U.S.S.R. signed a bilateral
agreement pledging cooperation on environmental protection. As a part of
this agreement, a working group on stationary source air pollution
control was formed by the U. S. Environmental Protection Aaency and the
U.S.S.R. Research Institute of Industrial and Sanitary Gas Cleaning to
define joint programs.

     The planned cooperative programs encompass several areas of air
pollution control technology, including particulate emission control.
High mass-collection efficiencies are now achieved on particulate
emissions from industrial processes in both countries by utilizing
electrostatic precipitators, fabric filters, wet scrubbers, and novel
devices.  Growing concern for the health and environmental effects of
fine particulate emissions (3 microns or smaller) has resulted in a need
for further improvement of conventional control techniques and for the
development of new techniques for fine particulate control.

     In order to exchange technology on fine particulate control, a
joint testing program was established.  Soviet specialists would visit
the U.S. to test with U.S. experts a hot-side electrostatic precipitator
and U.S. experts would join Soviets to test a high efficiency scrubber
in the U.S.S.R.  This paper discusses the second part of that program.
The high efficiency scrubber selected was a high pressure drop venturi
scrubber on an electric arc furnace at the Nikopol Ferroalloy Plant in
the Ukrainian Soviet Socialist Republic.
                                  210

-------
Description of the Test Site and Scrubber

     The Nikopol Ferroalloy Plant produces silicomanganese with electric
arc furnaces each of which consists of a rectangular bath and six oval
electrodes.  The electrodes are supplied power by three 21,000 KVA trans-
formers.  The rectangular bath is 20.3 m long by 6.0 m wide by 3.2 m
high and has three tapholes.  The furnaces operate continuously except
for scheduled outages averaging one day per month.

     The gases leaving the furnace are primarily carbon monoxide
(64-88%) and carbon dioxide (7.5-24.0%).  Other constituents are shown
in Table 1.  The gases contain as much as 20 g/Nm  of particulate matter
which is primarily silica (12-15%) and manganese oxide (10-16%).  The
particulate size is initially small; that is, substantially less than
1 ym.  However, the particles agglomerate to shift the mass median diameter
to as high as 3 pm.

     The particulate control system consists of a series of scrubbers
shown in Figure 1.  The first scrubber, which is item 1 in Figure 1,
is an inclined gas line (500 mm in diameter) into which water is sprayed
by two centrifugal injectors.  The second scrubber (item 2) is a spray
chamber (1000 mm in diameter) which contacts the gas countercurrently.
The third scrubber (item 3) is a venturi scrubber (100 mm in diameter)
with a pressure drop of 22 to 24 kPa.  The venturi scrubber is followed
by a cyclone mist eliminator (item 4).  The scrubbing water flow rate
for each injector in the inclined scrubber and for the venturi scrubber
is 7 m /hr.  Also shown in Figure 1 are the exhaust draft fan (item 5),
the filter press (item 6), the liquor recycle pump (item 7), the
sedimentation tank (item 8), and the cooling tower (item 9).

Experimental

     Measurements to determine performance of the venturi scrubber were
made before the venturi and after the mist eliminator.  Figure 2 shows
the location of the sampling ports.  Because sampling ports were not
available in the gas line before the venturi, ports in the spray chamber

                                   211

-------
scrubber were used for inlet sampling and the water to the spray
chamber scrubber was turned off.   Outlet sampling was conducted on a
long run of gas line after the mist eliminator.

     Mass concentrations and mass emission rates were determined using
an EPA Method 5 probe.  However,  only single point sampling was used
because of the small duct sizes and because of the problems in maintaining
gas tight seals to prevent combustion of the carbon monoxide rich flue
gas.  Particle size measurements  were made using cascade impactors:
Brink at the inlet; and Andersen  at the outlet.   Both impactors used
glass fiber collection substrates for the stages and a glass fiber final
filter for collecting particles which penetrated the impaction stages.

Results and Discussion

     Particulate concentrations into and out of the venturi scrubber
are shown in Table 2.  The average inlet concentration is 5.19 g/Nm ;
                  3
outlet, 5.94 mg/Nm  and the averaae efficiency is 99.89 percent.  Results
of particulate sizing data give a log normal size distribution with an
average mass median diameter (MMD) at the inlet of 1.7 ym with a geometric
deviation (ag) of approximately 2.  At the outlet the MMD is 0.42 ym
and the ag is between 2 and 3.  The fractional efficiencies are given
in Table 3.  The collection efficiency for particles areater than 1 vm
in diameter is greater than 99.9  percent.  As the particle diameter
decreases to 0.3 ym the efficiency decreases to 98.6 percent.

          The total penetration of the scrubber may be predicted by
equation 8.2-12 of the Scrubber Handbook^  .  This equation combines
an assumed log normal particle size distribution with the equation for
penetration of a monodispersed aerosol through a venturi scrubber.
Equation 8.2-12 can be solved by  integrating with respect to particle
size.  This integration was made  using the extremes in gas flow rate
which gave the extremes in throat velocity and hence the extremes in
total penetration.  For a flow rate of 2180 Nm /hr, the throat velocity
is 76.04 m/s and the predicted efficiency if 99.84 percent.  For a flow
                                    212

-------
               3
rate of 2500 Nm /hr, the  throat velocity  is 88.42 m/s and the predicted
efficiency is 99.94 percent.  The predicted overall efficiencies  bracket
the observed overall efficiency of 99.89  percent.

     With respect to fractional efficiencies, the efficiency for  a
single particle size may  be calculated from equation 5.3.6-5 of the
Scrubber Handbook.  The predicted values  from this equation are greater
than 99.9 percent for all particle sizes  noted in Table 3 except  for 0.3
urn. At this smallest particle size, the predicted efficiency ranges from
93.8 to 98.2 percent depending on the velocity in the throat used in the
calculations.  When all experimental and  predicted fractional efficiencies
are compared, the predicted values differ from experimental ones  by less
than 0.7 percentage units and are higher  for all particle sizes except
for 0.3 ym where the predicted value is lower.

Conclusions

     A high pressure drop venturi scrubber at the Nikopol Ferroalloy
Plant (USSR) was tested for total and fractional collection efficiency
of electric arc fume.   At a pressure drop of 22 to 24 kPa, the overall
efficiency of the scrubber ranged from 99.86 percent to 99.92 percent.
The collection efficiency for particles larger than 1 ym in diameter
was greater than 99.9 percent and for those smaller than 1 ym was areater
than 98.6 percent.  These experimental results are consistent with
predicted values from theoretical equations.

Acknowledgements

     The author wishes to acknowledge the work of Soviet specialists and
technicians who used the U.  S. equipment to produce the raw data  from
these tests.   Moreover, Mr.  Bruce Harris of the U. S. Environmental
Protection Agency and Mr.  Joseph McCain of Southern Research Institute
supervised the taking and reduction of the measurement data at the
Nikopol  Ferroalloy Plant.   The SR-52 program for integration of the
scrubber equations to calculate all penetrations was provided by
Dr. Leslie Sparks of the U.  S. Environmental Protection Agency.

                                  213

-------
  Table 1.   Properties of Nikopol  Electric Arc Flue Gas
Composition by volume percent
CO
09
2
C00
L
H2
CH, and N0
64-88
0.4-1

7.5-24.0

3.0-6.0
<8
Dry gas density - 1.3 kg/Mm


Heat value      - 9000 KJ/Nm3


Water vapor content - less than 20%

                                         3
Particulate concentration - 15 to 20 g/Mm


Particulate composition in percent of mass


               S^g                12-15


               MnO                 10-16


               CaO                  4-5


               NiO                  3-4


               A1203                1-3


               C                    7-10
                           214

-------
     Table 2.   Results of Concentrations Measurements at Inlet
               and Outlet of Nikopol  Scrubber
                                                       Efficiency
Day
(July)
28
29
30
30
31
31
Inlet
g/Nm
4.37
5.69
4.76
3.72
_ _ -
7.42
Outle^
mci/Nnf
3.55
5.35
6.49
4.77
5.85
9.13
                                                       99.92

                                                       99.90

                                                       99.86

                                                       99.87



                                                       99.88


Overall         5.19                5.94                 99.89
                                  215

-------
          Table 3.   Nikopol  Scrubber Fractional  Efficiency
Aerodynamic
Di ameter
   5.0

   3.0

   2.0

   1.3

   0.8

   0.5

   0.3
  Inlet
Incremental
Concentration
  mg/m	

  1500

  5700

  7400

  6100

  1050

   400

   280
   Outlet
Incremental
Concentration
    mg/m	

    0.85

    0.59

    1.14

    1.29

    1.63

    2.60

    3.90
Collection
Efficiency
   99.94

   99.99

   99.98

   99.98

   99.84

   99.35

   98.61
                                  216

-------
CO
D

<
DC
<
Q.
Q.
CD
QQ
2
o
o>
                                                 217

-------
O

4J
to

-------
Reference

     1.  S. Calvert, J.  Goldshmid,  D.  Leith,  and  D.  Mehta,  Scrubber
Handbook. EPA Report No.  EPA-R2-72-118a,  NTIS No.  PB-213-016,  August  1972.
                                  219

-------
                   A.P.T. FIELD EVALUATION
                  OF FINE PARTICLE SCRUBBERS
                       Seymour Calvert
                       Shui-Chow Yung
                       Harry Barbarika
                        Gary Monahan
               Air Pollution Technology, Inc,
                 San Diego, California 92117
                      Leslie E. Sparks
                       Dale L. Harmon
            U.S. Environmental Protection Agency
       Industrial  Environmental Research  Laboratory
          Research Triangle  Park,  North Carolina
                          ABSTRACT

     The performance of thirteen industrial scrubber systems
have been determined experimentally.  The performances are
reported in terms of grade penetration curve and cut/power
relation.
     The measured performances of these scrubbers are compared
with predictions of design equations.  In cases where design
equations are not available, they are developed from basic
theories.  Except for the design equation for the mobile bed
scrubber, design equations for other types of scrubbers tested
give reasonable performance predictions.
     The scrubber performance, expressed as the cut diameter
versus power relationship developed by Calvert has been veri-
fied and extended with the performance data.
                             221

-------
                        INTRODUCTION

     The need for more reliable data on the fine particle
collection efficiency of air pollution control scrubbers has
become increasingly apparent as control requirements have grown
more demanding.   Major efforts, such as the Wet Scrubber System
Study (Calvert et al.,1972), which resulted in the issuance
of the "Scrubber Handbook", have augmented our ability to design
better scrubbers and to predict their performance.  Design
methods, including mathematical models, have been developed
from basic theory plus whatever good data were available, but
to a large extent they were untested.  Thus, one could not con-
fidently predict performance for present scrubber designs and
operating conditions or extrapolate into better combinations
of design and operation.
     It is very difficult to compare scrubber performances in
different situations without knowing efficiency as a function
of particle size, commonly called "grade efficiency".  Even on
an empirical basis, there had been so few carefully and pro-
perly conducted performance tests that the capabilities of exist-
ting systems were not known.  Because of a predominant concern
for only the outlet particulate loading or emission rate, col-
lection efficiency in terms of overall particle mass was rarely
tested.  The few data which had been published were generally
unsatisfactory for use because of inadequate methodology, unde-
fined parameters, insufficient quantity, and similar inadequacies
     The program reported here was supported by EPA over the past
5 years and was initiated in response to the need for additional
reliable performance data on fine particle scrubbers.  The
objectives were to obtain data on fine particle collection
efficiency as a function of particle size for scrubbers opera-
ting on representative industrial emission sources and to recon-
cile the performance data with existing mathematical models.
                              222

-------
FIELD SAMPLING METHOD
     The method of approach to the program objectives involved
a number of experimental determinations to obtain collection
efficiency data, the acquisition of information on system char-
acteristics and behavior, and computations which utilized the
performance data and mathematical models.  Over the course of
the program the methods and apparatus used were generally im-
proved and were modified to suit each specific test situation,
but the main features were similar and will be described here.
     The most important experimental measurements were those
regarding particle size and concentration.  In the beginning
of the program the size range of primary interest was from a
few tenths to a few microns diameter, which is within the meas-
urement range of a cascade impactor.  Later the size range was
extended downward by an order of magnitude and it was necessary
to use a diffusion battery in addition to the cascade impactor.
The apparatus and methods used are outlined below:
     1.  Gas velocity distribution and parameters had to be
measured at the inlet and outlet of the scrubber in order to
define the following:
     a. conditions for isokinetic sampling,
     b. particle concentration per unit volume of dry gas, and
     c. gas flow rate.

Methods to measure these parameters are:
Parameter
                      Equipment
    Method
Gas velocity
and flow rate
                Standard pitot tube or cali-
                brated type "S" pitot tube;
                differential pressure gauge.
Gas temperature  Calibrated thermocouple or
                mercury filled glass-bulb
                thermometer.
EPA method 1;
EPA method 2.
Humidity
                Thermometers.
Pressure
                Inclined water manometer
                or a pressure gauge.
Wet and dry bulb
temperature measure-
ment on a flowing
sample withdrawn
from the duct.
Measured by means of
a static pressure
tube inserted in the
duct.
                              223

-------
     2.  The most essential part of the scrubber performance
tests is the determination of particle size distribution and
concentration (loading) in the inlet and outlet of the scrubber.
For accurate determination of particle size distribution, a
mechanism that collects particles and causes neither formation
nor break up of aggregates is necessary.  Cascade impactors
come close to meeting these requirements.
     In a cascade impactor, particles  are  classified  by inertial
impaction according to their mass. The larger ones are collected
on the plate opposite the first stage and the smallest on the
plate opposite the last stage.  A.P.T. uses Brink, Andersen
University of Washington Mark III, and a cascade impactor of
A.P.T.'s own design for particle size fractionation.   These
impactors (except Brink) classify particles into seven size
groups and are capable of sizing particles down to about 0.1 ym
diameter.
     In order to minimize probe losses all tests were made with
the impactors in the duct and with the inlet nozzles appropri-
ately sized to give isokinetic sampling.  A modified EPA Method 5
train was used to monitor the sample gas flow rate.
     In some tests, a pre-cutter was used to remove either the
heavy particle loading from inlet samples or the entrained liq-
uid from outlet samples.  A cyclone separator with about a 3 ymA
cut diameter was first used but a round jet impactor with about an
8 ymA cut diameter was found to have better characteristics and
was adopted for use for both inlet and outlet sampling.   The im-
pactors were either given time to reach the duct gas temperature
or heated to prevent moisture deposition.
     To increase the weighing accuracy, each impactor collection
plate was lined with light-weight substrate. Generally,  either
greased aluminum foil or a glass fiber filter paper  substrate
was used.  Impactor substrates and backup filters were weighed
with an analytical balance to the nearest tenth milligram (10"l+g).
     Particle size distribution and loading measurements were
conducted simultaneously at the scrubber inlet and outlet.  The
method minimizes the effects of particle size distribution changes

                             224

-------
caused by fluctuations in the operation parameters.  Since the
program objective was to investigate scrubber performance on
fine particles, the sampler was held at one location in the duct
for the duration of each sampling run.  This is an adequate tech-
nique for obtaining good samples of particles smaller than a few
microns in diameter because they generally are well distributed
across the duct.

RESULTS
     A summary of the scrubbers tested is given in Table I. The
first ten scrubbers in Table I are classified as conventional
scrubbers.  The last three are novel devices because they are
different in some manner from conventional technology.
     The dashed lines in Figures 1 through 10 are experimental
grade penetration curves.  They were computed from simultaneous
inlet and outlet particle size/concentration data.   The penetra-
tions were obtained from a plot of the cumulative particle mass
concentration versus particle aerodynamic diameter.  We use the
symbol "ymA" for aerodynamic diameter, which is equal to parti-
cle diameter (d ) in ym (microns)  times the square root of the
particle density (pp) in g/cm3 times the square root of the
Cunningham slip correction factor (C?).  The computation of pen-
etration (penetration is equal to  1  minus efficiency) as a
function of particle diameter is based on the following equations.
     Overall penetration can be defined as:

                  Pt = ~-  f P  (Pt)d  dCp                  (1)
                        T) L  J          U
                             o

     Penetration for particles with diameter dp  is given by:
                                                             (2)
f(dp)outlet
f (dp)inlet
" dCp "

d(dp)Joutlet
" dCP 1
d(dp)
inlet
                              225

-------
     " dCp/d(dp) " is the slope of the cumulative mass versus
aerodynamic diameter plot.  It may be obtained either by a step-
wise graphical procedure or by fitting a mathematical function
to the data point and then evaluating the slope analytically.

DISCUSSION
     Comparison of the experimental results with mathematical
models was done wherever models were available.  Table II is a
list of the design equations taken from the "Scrubber  Handbook".
All these equations are based on particle collection by inertial
impaction.
     The predicted grade penetration curves are represented by
the solid lines in Figures 1 through 10.  By comparing the pre-
dicted and the measured grade penetration curves, we can either
verify or reject the design equations.  The results of this
comparison are summarized as follows:
  1. Valve tray on urea prilling tower - No specific perfor-
     mance model was available for a valve tray so it was
     necessary to develop one.  We used the model for sieve
     plates because the gas jets emerging from the slots be-
     tween the valve cap and the tray impinge on a liquid
     froth similar to the round gas jets on a sieve plate.
     The model compared well with the data after accounting
     for particle growth due to water vapor condensation.
  2. Vaned centrifugal on KC1 dryer - The collection effi-
     ciency of this scrubber could not be accounted for
     simply by centrifugal deposition caused by the internal
     vanes.  A gas atomized spray model gave predictions
     which agreed with the data.
  3. Mobile bed on coal-fired boiler - No satisfactory model
     is available.  The mobile bed design equation is a semi-
     empirical correlation proposed by Bechtel based on the
     data reported by UOP.  It was developed on the assump-
     tion that collection efficiency is due to inertial im-
     paction on the balls.
  4. Venturi on coal-fired boiler - The model for a venturi
     in terms of particle cut diameter correlated with
     pressure drop agrees well with the experimental results.

                              226

-------
 5. Wetted fiber on Nad dryer - Heated impactors were used
    in the field sampling.  Thus, the measured particle size
    is that of a dry particle.  In the scrubber, dry particle
    size is not the actual size the scrubber will see.  The
    scrubber sees a wet particle which is larger.  We
    did not measure the wet particle size in the field.
    Common salt particles are highly hygroscopic.  According
    to Junge (1963) the physical diameter of the particle
    will increase to about   5  times that of the dry salt
    particle at high humidity and about double at 75% rela-
    tive humidity.
    Wet particle diameter was calculated from dry particle
    diameter with the assumption that its physical diameter
    is doubled.  Calculation results are shown in Figure 5.
    The filters in the filter pad were ellipsoidal in shape
    with the longer axis normal to the direction of gas flow.
    Therefore, its collection efficiency should  lie somewhere
    between the collection efficiencies  of  a ribbon and a
    cylinder.
    It can be  seen that the model agrees with the calcula-
    tions for  wet particles.
 6. Impingement plate on NaCl dryer - A model based on im-
    pingement  from round jets gives good agreement with the
    data after allowing for particle growth due to condensa-
    tion.
 7. Venturi rod on cupola - The venturi model gives a good
    prediction for particles larger than about 1.0 ymA but
    does not account for low penetration for the sub-micron
    particles.
 8. Venturi on asphalt dryer - The venturi model prediction
    agrees with the performance data.
 9. Venturi on borax fusing furnace - The model agrees with
    data for particles with diameter larger than 1 ymA.  For
    particles  smaller than 1 ymA diameter,  the model predicted
    penetration a few percent higher than measured.
10. Variable rod on cupola - The model predicted penetration
    higher than that measured.
                            227

-------
CUT/POWER RELATIONSHIP
     When scrubbers are operated at different pressure drops
it is very difficult to evaluate and to compare their perfor-
mances based only on grade penetration curves. We have developed
a useful correlation called the cut/power relationship for this
purpose and others. The cut/power relationship is a plot of the
cut diameter given by the scrubber against pressure drop or
power input, as illustrated in Figure 11. Cut diameter is the
particle diameter whose collection efficiency is 50%.  The solid
lines in this graph were calculated theoretically from the de-
sign equations presented in the "Scrubber Handbook."
     The performance cut diameters determined experimentally
in this program were plotted against measured pressure drop in
Figure 11.  (In cases where penetration curves do not reach 50%,
the reported cut diameters were equivalent cut diameters cal-
culated by the method presented by Calvert, 1974). As can be
seen, curves for impingement and sieve agree with the data. The
line for venturi scrubber from the previous correlation slightly
overestimates the pressure drop. The dashed line fits the experi-
mental data determined in this study and is based on a re-
vised method we have developed for predicting venturi scrubber
pressure drop (ref:   Yung  et al., 1976).
NOVEL DEVICES
     Under the novel device test program, we have tested a
CHEAP, an electrostatic scrubber, and a charged droplet scrubber.
Electrostatically Augmented Scrubber
     The electrostatic scrubber is designed by Air Pollution
Systems.  It is essentially a venturi scrubber with a charging
electrode placed ahead of the throat.  The unit we tested is a
pilot scale unit with a capacity of about 28 m3/min (1,000 CFM) .
The experimental data are shown in Figure 12 for the scrubber
system with the charger off and with the charger on.
     The scrubber manufacturer did not have a design model so
we had to develop one.  With minor modifications, the design
equations for the venturi scrubber apply.  In a venturi scrubber

                             228

-------
the most important unit mechanism responsible for particle col-
lection is the collection by drops and the predominant collec-
tion phenomenon is inertial impaction.  When particles are
charged, then, in addition to the inertial force, electrostatic
forces are present that force the particle towards the drop;
i.e., increase the collection efficiency of the drop.
     Calvert et al. (1973) calculated the theoretical total collec-
tion efficiency due to particle deposition caused by flux
forces plus inertial force.  A plot of single drop collection
efficiency against inertial impaction parameter with flux
deposition number, NpD, as the parameter is presented in Figure
13.  Flux deposition number is defined as:

           N   = _J1 = Particle deposition velocity          ^
            FD   U0   Gas velocity past drop                l

Assuming Stokes' law holds, the particle deposition velocity is
given by:                 C' 0  E
     Single drop collection efficiency, n, is related to scrub-
ber penetration by an equation given by Calvert for a venturi:
                       2 QT PT d, ur   f°
             (Pt) ,    = - L  L  d  Gt  I  n df              (5)
                 dp       55 QG ^G     ^a

     This model was applied to predict the collection efficiency
of the A.P.S. scrubber both with the charger on and with the
charger off.  Figure 12 shows the predicted A.P.S. scrubber  per-
formance along with experimental curves.  As can be seen, the
predictions and data cross each other for both conditions.   The
design equations predict a higher penetration than actually
measured in the sub-micron region.
Wetted Fibrous Bed
     The CHEAP system is primarily a wetted fibrous bed scrub-
ber system.  It consists of water sprays to wet and clean the

                             229

-------
filter medium, a rotary drum containing a fibrous "sponge" fil-
ter medium, and a water bath reservoir for cleaning the rotary
filter.  Particle collection is due to filtration by the rotary
filter.
     The CHEAP unit we tested was on a diatomaceous earth dryer
exhaust.  The experimental results are shown in Figure 14. The
grade penetration curve for wet particles in the figure was
calculated from the penetration curve for dry particles and par-
ticle growth data.
     The prediction of particle penetration for the CHEAP re-
quired the development of mathematical models because none were
available from the literature.  We can mathematically represent
the filter as a fiber bed consisting of an array of equally
spaced cylinders.  The "Scrubber Handbook" gave the following
equation for the prediction of particle penetration of a clean
fibrous bed:
(Pt),   =  exp   - S n   =  exp
                                      4 A(l-e)
                                        TT df
(6)
where "n" is the effective collection efficiency of a single
fiber in the bed by all collection mechanisms.
     By assuming negligible interaction among fibers, the col-
lection efficiency of a fiber in a fiber bed can be approxi-
mated by the collection efficiency of an isolated fiber.  This
assumption might slightly underestimate  the fiber efficiency
but, since the fibers are not all oriented normally to the gas flow
direction, collection efficiency will be lowered slightly.
     The pressure drop across the fiber bed is the sum of the
drag losses of all fibers.  We used the drag coefficient for
an isolated cylinder, to be consistent with the assumption of
negligible interaction among fibers.
               AP = 6.5 x10"
                                            •G
                                                             (7)
                             230

-------
     Equations  (6) and  (7J were used to predict the performance
of the CHEAP.   The result is shown in Figure 15, a plot of cut
diameter versus pressure drop for various fiber diameters.  The
cut diameter versus pressure drop relation is highly dependent
on the diameter of the  fiber, but not much on the solidity fac-
tor.  The circle in the figure represents the data we  deter-
mined experimentally.   The fiber diameter and the porosity of
the filter medium of the scrubber were not disclosed to us be-
cause they were proprietary data.  Therefore, we could not
determine how well the  model predicts the performance.
     It is possible to  compare our model with the data of Rei
and Cooper (1976) for tests on a pilot scale unit of the CHEAP.
They reported the volume fraction void of the filter medium to
be 971 and the  fiber diameters to be 64 ym, 44 ym, and 36 vim
for foams with  18, 26,  and 33 pores per cm.  They also reported
the measured cut diameter was somewhere below 0.5 ymA for pres-
sure drops  ranging from 40 to 90 cm W.C.  The dashed line in
Figure 15 shows their data, which are  consistent with our pre-
dictions .
Charged Droplet Scrubber
     The charged droplet scrubber was developed by TRW Systems.
Instead of charging the particles, as in the case of APS elec-
trostatic scrubbers, the TRW charged droplet scrubber charges
the water drops.  The water flows out of small diameter tubes
which also act as electrodes.  The water is atomized as it jets
from the tubes.   Particle collection of this scrubber is due to
inertial impaction and the electrostatic force that exists be-
tween the particle and the water drop.   The 680 m3/min (24,000
CFM)  unit we tested was used to control emissions from a coke
oven.
     Figure 16 shows the experimental data.   A mathematical
model for the charged droplet scrubber was developed by TRW.
We did not compare the model prediction with data in this study.
                             231

-------
CONCLUSIONS
     In conclusion, the field sampling program reported here
has attained its major objectives.  We obtained meaningful per-
formance data and have verified or developed several mathemati-
cal models.  Except for the design equation for the mobile bed
scrubber, design equations for the scrubber types tested in
this program give reasonable performance predictions.  In cases
where "Pt" was lower than predicted it could be accounted for
by particle growth due to water condensation.
     The cut/power relationship has many useful applications.
It can be used to compare and evaluate scrubbers, to make pre-
liminary scrubber selections, and to estimate the minimum pressure
drop of a scrubber for it to attain the required performance.
We have verified and extended this relationship in this study.
ACKNOWLEDGEMENT
   The work upon which this paper is based was performed
pursuant to Contracts 68-02-0285, 68-02-1328, 68-02-1496,
and 68-02-1869 with the Environmental Protection Agency.
                             232

-------
                      NOMENCLATURE
    C~ = drag coefficient,  dimensionless



    C  = particle concentration,  g/cm3



   C   = total particle concentration, g/cm3



    C' = Cunningham slip factor,  dimensionless



    d  = collector diameter,  cm



    d, = drop diameter, cm



    dr = fiber diameter, cm



    d,  = diameter of perforation, cm



    d  = particle diameter, ym



   d   = aerodynamic particle diameter,  ymA
    pa


   d   = mass median particle diameter,  ymA
    ± &


     E = field strength, kV/cm



     F = foam density,  dimensionless



     f = empirical constant,  dimensionless



    f  = empirical constant = 0.5
     a


 f (d ) = frequency distribution of particles



    K  = inertial impaction parameter, dimensionless



   Kp  = inertial parameter in the venturi  throat,

         dimensionless



     £ = thickness of filter  pad, cm



   NpD = flux force deposition number, dimensionless



     n = number of stages,  dimensionless



    Ft = overall penetration,  fraction or percent



(Pt) i   = penetration for particles with  diameter d  ,

     p   fraction                                P
                           233

-------
             NOMENCLATURE (continued)


 Qg = gas volumetric flow rate, m3/s

 QL = liquid volumetric flow rate, m3/s or H/s

 Q  = electrical charge carried by the particle,
  "   coulomb

  S = solidity factor, dimensionless

 Tp = gas temperature, °C

 Up = particle drift velocity, cm/s

 Up = gas velocity, cm/s

Up. = interstitial gas velocity, cm/s

Up. = gas velocity in the venturi throat, cm/s

 u, = velocity of gas through perforation, cm/s

 u  = gas velocity, past drop, cm/s

  z = static bed height, cm


Greek

  e = porosity, fraction

  i = filter pad thickness,  cm

  r) = single drop or single fiber collection efficiency,
      fraction

 Up = gas viscosity, poise

 Pp = gas density, g/cm3

 PL = liquid density, g/cm3

 p  = particle density, g/cm3

 a  = geometric standard deviation

 Ap = pressure drop, cm W.C.
                        234

-------




























o
w
uu
co
LU
CO
CO
Q*
O
CO

p-
o

o:
$
y
»-^
CO


l-H

IU
r-H
fy
CD
H






















i rt
rH IU •-
O O C
4-i C
i- —
0 <
M E
rH TO 4^ ID P
0 E 3 E ^-
p U
4




t
4-1 TH C
GJ O rH O
rH t) 4-J -r-
C -H '-T, 4-J
l-H CO 'i-l 3
Q X!


j

_)
- ='

o

-1

_f
LO

ac LO
D

<;


1—1


p. 1
-1
M
H -1
•01
o
CU S-, 0~ 3
r-. 3 !•
O. 10 C


u
*
.
(_
E—


• G

C^
to

O
4-.' i--
3
O

3 C
— <
I—1

r^
CM


<^V

O/




cj;
o-






































B
erf
^
0


c

'E

Crt
1
VO
00







bo
c
o
rH
3
o
CO
• H
l-H

• rH
rH rH
D. CU

CD O
CU 4-J
rH
3
X
0)
rH

• r-t
CD
CJ
•H
£
o
1
o

4->
c
o
c_>
X
H
l|H rH
• H

rt 3
O
U






f—^
^
^
(X
o
r>

•8
X)

CU
•H
rO
o


to

^
vD

C

^
CM



LO



00

to


LO
(N


^r
tn

t^
sC
r— \

CO
rH


0
0


to
i-H

0>
l-H
II-H
O
X)
•o
CU X

•H >H
C4H rH


t


*G

C

r--
CM

00

•*





0
i-H

CTl
l-H


CM
to

00
to


CM
0


o
CTl
LO

l-H






rH
CU
X
rH
•a

4_j
rH
M
CO







rH
CU
4-J
r-H
•H
<4H

J_j
X)
•H
^M

"S
^_)
CU
•s

LO


to

-1

C
rH


1



O
0
rH
A


O
to


oo
to

LO
00


l-H
c


00
to
CM








L,
CU
X
rH


+-*
l-H
cd
CO








0)
,-% rt +J
p, i— 1 Q)
$-t QH 'f— »
o c
" g'l
rH 0 r-H
O 

!-•

^
to

O

to
o



to





o
i-H

vO
'-O



LO

Ol

1— 1

to
•-H


o
o
00






_f*
o
cd
XI

4-* 4-*
rH C
Clj rt
_C i— i
o. a

"*










^— ^
U-


• H
3
4_)
CU


00


CM

o

LO
CN



to





r^


O
l-H


LO
LO

O
00


r^
i-H


O
o
CM

i-H


aO CU
U
rH 0)
CU C
C M
•H 3
U 4-1
l-H
a) bo
o C
•H
X in
rt 3
rH Ci
o
pa










^—^
p.
^ — '

• H
IH
•P
C
CU


0>

00


0

LO
i-H

00

l-H



LO




o
CO
I-H

O
CM

LO
to


l-H
CN


O
l-H
O

I-H



nl
l-H
O
(X
3
U

X

T3
C
3
O
0-
-£•
O

o
CU
l-H
•H l-H
rH O
3 U
4-"
S S
> 3
Q
•a
O i-H
M CD
I C
CU O
I-H 'H
XI 4J
rt a)
•H 2
rH v— '
CD


0
l-H


00

o


LO



•*



LO
00



to
LO

OO

O


to



LO
r-
o


o
i-H
[ —



-C
n)
CU

to
o

CJ
CD
E
O
4-1
a)
•H
Q





CU
l-H
l-H
• H
C
nl

i
to
C
•g
^"3
*-_-'
u.

1
CJ

t-H









































rH
&
•a

ufl1
L,
CU
c
• H
0
l-H
CD
O
























LO
to

0

LO
oo

l-H

CM





1


0
<--J-


VO
i-H

\o
rH


CO
rH


to
(N






4J
Ul
3
T3
4_>
(A
O
+J

fsj
O
• H
E-


w
e
rH S
CU 4-J
XI V)
XI X
3CO
0 C
to O
• H
CJ 4->
•H 3
4J l-H
(D rH
4-1 O
in a.
O
4-1 «H
0 <
(U »_/
r-H
LU

CM
rH

LT
to

O

01
00
OO
,_,

CM



*3-

j


O
l-H


0


o
CM
• — !

LO
0
o


0
00
vO









c
CU
>
o

o
££.
o
u









CU
i-H
PH
0
rH
•a

'O ^^
(P ^
bo a:
rH H
KJ ^
^
a

fO
rH























































4
T-H
0)

s
W)
•rH
U*

p;
"S

• H

rH
d>

1


CD




-------


CO
s
H
t;
PQ
PQ
D
U
CO

r^>
£3
O
1 — 1
&
<<
£>

oi
^,
IX

t/:
^
0
*—
:— •


II

P.
p^

I-O
vC
•
0

1 1
p.
X UH
^
J_
C 00
"^" hO
t •
| ^

i

r i

O
X)









w
1 p











)>*
£-H

(X
W
CQ
M
Oi
U
CO



X
0
II

P-
t?
OH
re
t-

u
p#
"H
CO
>
•a
c
s
4)
O
P-»
CO

u
u T3

^ P,
^


^j
ID
m
^ "^
'^
a.

U
3


f
"y>-^*™^^^^ "H
Jl C; O U
c/ Id- 3 -a

•^ ^ u
cr as ~i
r~ P,
x "= ! a'
u". II
•
C' a
i ^
1 	 1
P.
X
(l)
II

p.
"^
*->
e-








"2
^
0
-rH
,£>
O
IS












1
1+




£a
'_
U


T3
T)

, •)
Q.

•t-*
3

J
cx

L1^

1

i OH

•^^ __


}
-<
a,
s
^
a.
+
'^
o


•—^



r--
.
o




^^

G
T—

«5

CZ3 t-
^1 i

cy3
rH
un
LO
1
,
H

1-
(
H
1

^
OH
y>



P, M-l
>
a
<
J
u


/-
•
c
s.



(/]
0)
0£

•a
s
rt

•H
3
C
O
>
CL "*
T3 ^"
-* U
-)
L,
^f




^.^
ro
^
P.

T3
CD
N
•tH
£
O
•M


"
C,
.*
*t









































	 v
Ijj
1
i— 1 (-M
• 	 i •— )
.
_ ^ ^
(./)
II
C
1 r^
-'
.0,
X
CJ
II

T3
/ — .
£
r-
CU
r— 1
•H
H-4

VH
(b
X)
•H
14-1
•u
0>
*->
fll
2





















Ul
o
^^HM*^^^^

m
f-
-a

c
u
.""i CJI
O*
r-~
f",
•
- — ^
II

m
rt
P.
T3








^J
C
E

-------
   1.0
o
• H
•P
u
cti
o
Pi
H
W
Z
W
PH
   0.1
PREDICTED



EXPERIMENTAL
  0.01
                        I
                          l  I i  i
                   1
     0.1
           1.0


PARTICLE DIAMETER,
                                           10
  Figure  1.   Predicted and experimental penetrations
             for Koch Flexitray.
                     237

-------
   1.0
ti
O
•H


2  °'5
2
O
E-
PJ
Z
tu
CU
   0.1
PREDICTED


•EXPERIMENTAL
      0.1
                   I   i  I i i  i i I
                                            1I|I  I I !_
                                    I  I  I I  I I
            0.5     1.0               5.0

            PARTICLE  DIAMETER,  ymA
10
    Figure 2.  Predicted and  experimental  penetrations
               for Ducon Multivane  scrubber.
                               238

-------
 1.0
                                      I    I   I  L I  I I I
0.01
    0.1
 0.5     1.0

PARTICLE DIAMETER,  ymA
5.0
  Figure 3.   Predicted and experimental penetration
             for mobile bed scrubber.
                         239

-------
   1.0
c
o
u
nJ
!H
4-1
2
O
i—i
H
<
PS
E-
W
2
W
                           I I I I I
                             \
0.1
  0.01
              PREDICTED

              EXPERIMENTAL
                                                  I I I I.
                                                   I  l  i i i
      0.1
                             1.0

                   PARTICLE  DIAMETER,  ymA
10
    Figure 4.  Predicted  and  experimental  penetrations
               for Chemico venturi.
                              240

-------
    1.0
O
•H
•P
U
cfl
S-i
M-l
H

£
w
is
w
    0.5
    0.1
   0.05
   0.01
                          RIBBON-
                          FIBER
       0.1
                 PREDICTED

                 •EXPERIMENTAL

                 - CALC.  FOR  WET
                  PARTICLES
                                          CYLINDER
                                          FIBER
                          I  l  I l I I
                                         I
I	I
I  i i l
                         0.5     1.0
                       PARTICLE  DIAMETER,  ymA
     5.0
    10
    Figure 5.  Predicted  and  experimental  penetrations
               for Encort  wetted  fiber  scrubber.
                              241

-------
   1.0
o
Oj
2  o.i
f-
oi
E-
PJ
Z
  0. 01
                         I  I I I I
                             \
              PREDICTED
       u 	 EXPERIMENTAL
                  I  I  I   I  I  I I I I.
                      I   l  I I I I I
                  I    I   I  I  l l  i i
      0. 1
         1.0

PARTICLE DIAMETER, ymA
10
    Figure 4.  Predicted and  experimental  penetrations
               for Chemico venturi.
                              240

-------
    1.0
    0.5
o
•H
-P
U
rt
O
H
W
)2
w
a,
    0.1
   0.05
   0.01
                                         CYLINDER
                                         FIBER
                         RIBBON
                         FIBER
                                             I   I  I  l  i
               — PREDICTED

               — EXPERIMENTAL


               	 CALC. FOR WET
                  PARTICLES
                I	l
                          I  l  I l
I
I   I   i  i i
       0.1
                         0.5     1.0

                      PARTICLE DIAMETER, ymA
         5.0    10
    Figure 5.  Predicted and experimental penetrations
               for Encort wetted fiber scrubber.
                              241

-------
   1.0
c
o
o
rt
2

0  0 1
I—I  u • '

H
H
W
  0. 01
              PREDICTED
       - 	 EXPERIMENTAL
                                          I  I   I  I  IT
                                            I  I  I  I  I 1 I
      0. 1
         1.0


PARTICLE DIAMETER, ymA
10
    Figure 4.   Predicted and experimental penetrations

               for Chemico venturi.
                             240

-------
     1.0
  o   0.1
  •H
  U
  03
  2
  O
  i — i
  H
  <
  Pi
     o.oi
    0.0011
        0.1
                                EXPERIMENTAL  _
                      I  . . ..I
         1.0

PARTICLE DIAMETER, ymA
                                             10
Figure 7.   Predicted and experimental grade pene-
           tration curves for venturi rod scrubber
                      243

-------
     1.0
              i   T
                        I I I
                                       I  I I I I L.
     0.1
o
•H
4->
U
Oj
!H
M-l
o
I—I
E-
CX
    0.01
   0.001
               PREDICTED
          	EXPERIMENTAL
                     t  j i i i i
       0.1                 1.0

               PARTICLE DIAMETER, ymA
                                            i i
10
   Figure  8.   Predicted and experimental penetrations
              for AAF Kinepactor 32.
                         244

-------
      1.0
      0.1
£
O
• H

•(->
O
03
  O
  I—I
  H
  <
  OS
  H
  W


  £  o.oi
    0.001
                   II I I I I I
                                 I    I
                                      1 '
                                PREDICTED
                            	 EXPERIMENTAL  j
               I   i  i  i  . . ..1
        0.1
                                     ,  I
                        1.0

               PARTICLE DIAMETER, ymA
                                             10
Figure 7.   Predicted and experimental grade pene-
           tration curves for venturi rod scrubber,
                      243

-------
     1.0
     0.1
o
rt
o
HH

E-
g   0.01
w
   0.001
              PREDICTED
              EXPERIMENTAL
                 i  i
                                   i	I
0.1                1.0

        PARTICLE DIAMETER, ymA
                                             10
   Figure  8.   Predicted  and  experimental  penetrations

              for  AAF  Kinepactor  32.
                         244

-------
        0.1
    o
   •H
   •P
    U
    oj
    JH
   M-i
   Z
   O
   t—i
   H
   <
   OS
   H
   W
   2:
   w
   ex
       0.01
      0.001
                PREDICTED
                •EXPERIMENTAL
                 l  I  I i
          0.3          1.0

            PARTICLE DIAMETER,  ymA
3.0
Figure 9.  Predicted and experimental  grade
           penetration curves  for  AAF  Kine-
           pactor 56.
                  245

-------
    0.1
   0.05
U
CO
E-
<
Cri
H
DH
   0.01-
  0.005
  O.OOli
      0. 1
      0.5     1.0

PARTICLE DIAMETER, ymA
 Figure 10.   Predicted  and  experimental  penetrations
              for  variable rod  scrubber.
                         246

-------
 c
 OJ
 s
 
 ft   0)
 g  -H
          G
          a;
     CO  >

                                                                                               U 
                                                                                               E ^
                                                                                               •H XI
                                                                                               !H 3
                                                                                               0) SH
                                                                                               ft U
                                                                                               X U}
                                                                                               0)
                                                                                                  I— I
                                                                                               13 rt
                                                                                               C JH
                                                                                               03 
-------
  1.0
o
•H
+->
U




 •\


o
I—I



OS

w
2
  0.1
 0.01
0.005
                                    I  T
         	CHARGER ON

         	CHARGER OFF
                               I     I   I  I   I I  I I
    0.2
                        1.0

                  PARTICLE DIAMETER,  ymA
10
Figure 12.  Predicted and experimental penetrations
            for APS electrostatic scrubber.
                        248

-------
                                                                        OS
                                                                       •M
                                                                       CD
                                                                       6
                                                                       03
                                                                       ^
                                                                       03
                                                                       03
                                                                       0)
                                                                       10
                                                                       !-<
                                                                       CD
                                                                       CD    •
                                                                      rH   J-
                                                                       W)  CD
                                                                       C  4-"
                                                                      •r-i   CD

                                                                           03

                                                                       O   03

                                                                       X
                                                                       u  to
                                                                       C  rt
                                                                      CD
                                                                      U
                                                                      CD
                                                                           CD
                                              o  o
 0
 >H

 3
 CXO
•H

UH
249

-------
     0.1
 c

•2   0.05
 •M
 u
 05
 i-
o
u— 1

<
e:
E-
UJ

D-
    0.01
             I    I   III!
       0. 3
               NXWET PARTICLE
DRY PARTICLE
             I   I   I  I  I  I
                      I	I
         1 .0

   PARTICLE DIAMETER,  ymA
4.0
Figure 14.  Experimental grade  penetration curve for CHEAP,
                             250

-------
   3.0
            I ' ' ' 'I
   1 .0
O
U
                     df=300 ym
              A.P.T. DATA

              RE I AND COOPER'S DATA
   0. 1
              i  i i i
                    I
                     I
                   10                        100

                     PRESSURE DROP, cm W.C.
                                        300
    Figure 15.
Cut diameter versus pressure drop for
fibrous bed.
                              251

-------
   1.0
 U
 03
    0.1
 OS
 H
 O,
   0.01
              I     I   I   I  I  I
J	I    1  I  I  I I  I
                               1.0

                       PARTICLE DIAMETER, ymA
                  10
Figure 16.  Experimental penetration curve for charged
            droplet scrubber
                           252

-------
                        REFERENCES
1. Calvert, S., J. Goldshmid, D. Leith,  and D.  Mehta.   "Wet
   Scrubber System Study, Volume I, Scrubber Handbook."
   EPA-R2-72-118a (NTIS PB 213-016), August 1972.

2. Calvert, S., J. Goldshmid, D. Leith,  and N.C.  Jhaveri,
   "Feasibility of Flux Force/Condensation Scrubbing for
   Fine Particulate Collection," EPA 650/2-73-036
   (NTIS PB 227-307), October 1973.

3. Calvert, S., "Engineering Design of Fine Particle Scrubbers,"
   J.  of A.P.C.A., 2£, No. 10, p.  929, October  1974.

4. Yung, S.,  S. Calvert, and H.  Barbarika, "Venturi Scrubber
   Performance Model," Final report to EPA, Contract  68-02-
   1328, Task 13  (1976)  (in printing).

5. Junge, C.E., "Air Chemistry and Radioactivity," Academic
   Press, 1963.

6. Rei, M.T., and D.W. Cooper, "Laboratory Evaluation  of the
   Cleanable  High Efficiency Air Filter  (CHEAP)", EPA-600/2-
   76-202  (NTIS PB 256-698/AS),  July 1976.
                             253

-------
   RESULTS OF FLUE GAS CHARACTERIZATION TESTING AT THE
           EPA ALKALI WET-SCRUBBING TEST FACILITY
                Richard G. Rhudy and Harlan N.  Head
                         Bechtel Corporation
                             ABSTRACT
A series of tests were conducted at the EPA Alkali Wet Scrubbing Test
Facility located at the  Shawnee  Steam  Plant,   Paducah,  Kentucky,
to measure mass loading and the size distribution of the particulates
emitted from the lime  and limestone wet scrubbers.  The  tests were
designed to determine the effects of  operating variables on emission
levels.    Two different scrubbing trains were  tested,  a TCA and  a
venturi followed by  a spray  tower.   Flue  gas feed for the unit was
supplied from Unit 10  of the Shawnee  Steam Plant  either  directly
from the boiler  or after the electrostatic precipitator.
Mass emissions  from the venturi/spray  tower system were  in  com-
pliance with the EPA  New Source Performance Standards while  the
mass emissions  from  the TCA system approached or exceeded  the
EPA standard for the conditions tested.  Mass penetration by particle
size,  measured from 0.3 to  10 microns, was generally higher  for
the TCA  system  than  for the  venturi/spray  tower  system over  the
range of variables tested.
Slurry entrainment was small.   It appeared primarily as particulate
emissions greater than 2 microns actual diameter.
                                  25!

-------
           RESULTS OF FLUE GAS CHARACTERIZATION TESTING AT THE
                 EPA ALKALI WET-SCRUBBING TEST FACILITY
Introduction
The EPA Alkali  Wet Scrubbing  Test Facility, located at the  Shawnee
Steam Plant,  Paducah, Kentucky,  is a lime/limestone throwaway pro-
cess demonstration facility operated  under EPA direction.   The pur-
pose of the test  facility is to demonstrate the commercial feasibility
of lime and limestone wet  scrubbing of coal fired power plant  exhaust
gases.   '  '    At the test facility,  a venturi followed by  a spray tower
and a TCA  (turbulent contact  absorber) are operated to  remove  both
particulates and  SO2 from flue gas generated by the coal fired boiler.
As a part  of  this project,  a  Flue Gas Characterization Program is
being conducted to answer  a number of questions  concerning the flue
gas emissions from the lime and limestone wet scrubbers.
First, mass loading data was needed to determine the conditions under
which EPA  New Source Performance Standards  for mass  emission
could be met.  Second,  a measurement of the  extent to which scrub-
bers generate particulates through scrubber  reaction-product entrain-
ment was desired, especially in the fine  ( < 2  M  )  particulate range.
And  third, because of concern over the health hazards of sulfates,^
it was desired  to  determine the  effect of scrubbers on sulfuric  acid
vapor (SO 3)  emissions and, using this data, to estimate  total sulfur
emis«ions (i.e., SO2, SO3,  and reaction-product sulfates).
An intensive testing period to explore  the  effect of major  operating
variables on each system has  been completed.  This paper  reports
the  results  only  of  the mass loading and size distribution tests.  SO 3
emission results will be reported  elsewhere. Routine  testing is being
continued to monitor long term  trends  as opposed to the periodic spot
checks made in the past.
Experimental
                         Test Program
The  test series on both the venturi/spray tower and TCA systems were
                                  256

-------
designed to determine the effect of major operating variables on mass
removal and particulate size distribution.
Variables investigated on the venturi/spray tower system were flue gas
rate, slurry rate, MgO  addition, venturi pressure drop,  mist elimi-
nator configuration, and percent solids recirculated. Variables investi-
gated on the TCA system were flue gas rate, liquor rate, MgO addition,
and mist  elimination wash  scheme. MgO is used as an additive in
lime/limestone systems  to increase sulfur dioxide removal.  The mist
eliminators were washed intermittently on top with fresh water (8-hour
sequential cycle with one of the six nozzles  activated for  4 minutes
every 80  minutes) -while three modes  of bottom wash were  evaluated
as follows:
           High frequency intermittent fresh water  (HI),  4  min/hr at
           1. 5 gpm/ft  , shut off during testing
           Low frequency intermittent fresh water (LI), 6 min/4 hr at
           1. 5 gpm/ft  ,  shut off during testing

           Continuous dilute
           on during testing
                                                    o
Continuous diluted clarified liquor (C),  at 0.4 gpm/ft , left
For  one run on each system, flue gas feed to the  scrubber was obtained
from downstream of the electrostatic precipitator (ESP) to check scrub
ber performance  on  flue  gas with low grain  loading.  Flue  gas feed
for all the other runs came from a takeoff just downstream of the boiler
air preheaters.  Listed in Table 1 are the run conditions for the venturi/
spray tower system tests, and in Table 2,  the conditions for the TCA
system tests.
                         Sampling Locations
Both the inlet  and outlet of each  scrubber were sampled.  The duct
work at these locations was forty inches in diameter.  Two eight-point
traverses,  90° apart, were performed for each mass loading and size
distribution test. The  inlet sample  ports were located ahead of the satu-
                                  257

-------
ration sprays at the scrubber inlet  where the gas temperature was ap-
proximately  300-330° F.  The outlet sample port was located after the
reheater at the scrubber outlet where the gas temperature was approxi-
mately 250°  F.   The sample ports were about two duct diameters up-
stream and six duct diameters downstream of the nearest obstruction.
                            Test Methods
The procedureused for measuring particulate mass loading was a modi-
fication of EPA  Method  Five.    A  Hi-Volume sampling train  manu-
factured by the Aerotherm  Division of Acurex Corporation was used.
The major differences in test conditions between sampling at Shawnee
and a typical Method Five application were (1)  increased  sampling flow
rate (2  cfm versus  0.75  cfm typical for EPA Method Five),  (2) filter
temperatures inthe350°F to 400° F range to prevent acid  (SO3) conden-
sation,  and (3) the first two impingers filled with NaCC>3 solution to
remove SC>2  and to provide  corrosion protection for the pump and dry
gas meter.
Particulate size  distributions were measured with out-of-stack heated
inertial impactors.   A Brink Model BMS-11 impactor was used for in-
let sampling and a Meteorology Research Inc.  (MRI) model 1502 im-
pactor was used for outlet sampling.
To  obtain  representative  size  distribution measurements, sampling
traverses  were made.  The flow rate through  the impactor was held
constant to ensure  constant impactor stage cut sizes.  The sampling
nozzle was chosen to provide average duct velocity at the nozzle  inlet.
Because of the uniformity of the duct velocity profile,  +.10 percent of
the isokinetic sampling rate  could  be  maintained  by sampling  at the
average duct velocity.
Because all size distribution measurements were taken out of stack to
facilitate temperature control,  a fraction  of the particulate was  de-
posited in the probe.   For the outlet size distributions,  the estimated
probe fallout  included a significant fraction  of  the particulate in the
size range of  the first three stages.  Thus, it was necessary to com-
                                  258

-------
bine the probe wash and the first three  stages  when computing grain
loadings as a function of particle size.  For the inlet measurements
no correction was required.
Results
                  Venturi/Spray Tower System
Included in Table 1 are the results of the mass loading tests performed
on the venturi/spray tower  system.
As  expected, the venturi operating conditions had the greatest effect on
particulate removal.  Average outlet mass loading ranged from 0.019
gr/dscf at a venturi pressure drop of 9  in. H2O  and 600  gpm slurry
rate (Run VFG-1C) to 0. 036 gr/dscf at ~ 3 in. H^O and 140 gpm slurry
rate (Run VFG-1P).    Even at a venturi pressure  drop of  3 in.  H2O,
the venturi/spray tower  system was  capable  of  meeting  EPA New
Source Performance Standards  (NSPS) for mass emissions from coal
fired power plants  (0.052  gr/dscf assuming  30  percent  excess air
in the flue gas).   However,  the EPA NSPS opacity requirement  will
probably  limit the maximum particulate concentration allowable for a
commercial size stack.
Future testing  will concentrate  on defining the relationships between
opacity and mass loading.  Preliminary results of the first of these
tests  on the TCA system are discussed under the TCA results.
Average outlet mass loading for the run with low inlet fly ash concen-
tration (Run VFG-1B, flue gas from after the electrostatic precipitator)
was 0.005 gr/dscf.   Calculations based on size distribution measure-
ments suggest there should be no problem with  opacity at this  level
and it is apparent that entrainment from the circulating scrubber slurry
must be less than this value.
                                 259

-------
To further quantify the amount of entrained  scrubber reaction products,
the particulate from one outlet-mass-loading  filter was analyzed for
each of Runs  VFG-1A,  VFG-1B,  and VFG-1C.  Table 3 presents the
results of wet chemical and semi-quantitative  spectrographic analyses
of these solids.   By assuming all the calcium and  magnesium in the
samples were present as hydrated reaction products (as noted in Table
3), the mass  loading of entrained scrubber  reaction products was es-
timated.  For  the VFG-1B  analysis,  the resultant value  was 0.002
gr/dscf  while  for the VFG-1A and  VFG-1C  analyses,  the value was
approximately 0.006  gr/dscf. However,  the  runs were made at the
same operating conditions  (see  Table  1),  and the entrainment values
should be the same.   The values  for  Runs VFG-1A and VFG-1C may
be high because a portion of the  calcium  collected  from  these  runs
existed in the fly ash  emitted from  the scrubber and  should  not have
been included  in  the reaction-product emission estimation. Scanning
electron microscope  (SEM)  photographs  of  the outlet-mass-loading
filter  indicated that the  emitted   solids were predominately fly  ash
particles.
Inlet and outlet grain loadings as a function of aerodynamic particle
size were calculated '  using mean values of all the measurements for
each run. These values  are plotted for  runs with high fly ash loading
in Figure 1 and for the run with low fly ash loading in  Figure 2.  The
outlet  particulate  size  distributions, like the  mass  loading results,
appeared to be fairly independent of  operating  variable level for the
runs with high fly  ash loadings.   The outlet size  distribution for the
run with low fly ash (Run VFG-lB)  exhibited  a similar shape but had
lower values reflecting the reduced grain loading.
Mass penetration  as a function  of  aerodynamic particle  size,  deter-
mined graphically from Figures 1  and 2, are plotted  in  Figure 3 for
all runs  with high fly ash loading  and in Figure 4 for  the  run with
low fly ash loading.    Penetration appeared  to  be fairly independent
of run variable  level.  Mass  penetration for the high fly ash loading
tests appeared to vary  from  about 20 to  50 percent  for 0. 1 micron
particles to  less  than 1  percent for particles larger than 5 microns.
For  the  tests with low fly ash loading,  mass penetration was the same
for  0. 1  micron particles but penetration  increased to approximately
five  percent for particles larger than 5 microns.
For the  Shawnee  particulate emissions,  a 2 micron actual diameter
                                  160

-------
corresponds to a 3. 1 micron aerodynamic diameter." Using this rela-
tionship,  the irnpactor data was split into mass emission greater than
and less  than 2 microns actual diameter.  Mass emission of parti-
cles less  than 2 microns actual diameter  from the venturi/spray tower
system  was  insensitive  to  run variable level,  averaging ~ 0.016
gr/dscf for  all  of the high  grain loading runs.  Emission of particles
greater than  2   microns averaged~0. 006 gr/dscf for all runs except
Run VFG-1A (MgO addition) and Run VFG-1P  (minimum venturi). The
data for  these runs  was not of sufficient quality to  estimate a  mass
emission  for the  greater than two micron diameter  particulate.  For
the low fly ash  run,  emissions less than 2 microns actual diameter
averaged  0.002  gr/dscf while emissions greater than  2 microns actual
diameter  averaged 0.003 gr/dscf.
SEM photographs  (see Figures 5 and 6) of the scrubber outlet impactor
particulate  from  Run VFG-1A  captured on the third stage (7.8 to 9.3
microns actual cut size) and  sixth  stage (1.1 to 1.2 micron actual cut
size)  revealed that  very little  emitted reaction products were visible
on the sixth  stage (deposits appeared to be all spherical fly ash) while
crystals of CaSC^.  1/2H2O (the rosettes) were visible on stage  three.
These SEM photographs  confirmed  the presence of scrubber  slurry
reaction products and indicated they were  concentrated in the  larger
size range.
                              TCA System
Presented in Table 2 are the results of the mass loading tests conducted
on the TCA system.   The average outlet mass loading ranged from
0. 042 to  0. 065 gr/dscf for  the  high inlet  fly ash loading tests and
was 0. 026 gr/dscf for the test with low fly ash loading.
The average mass emission values with high inlet fly ash loading were
in the range  of and  sometimes exceeded the  EPA New  Source Perfor-
mance Standards for  particulate emissions.   Preliminary results  of
mass loading versus opacity measurements indicate that removals in
the 0.020 to 0.025  gr/dscf range would be  required to meet the EPA
opacity standard for a  500 MW plant with a single  stack (assuming a
70 ft/sec exit velocity).   This value is well below the mass emission
requirements.  Further  tests are planned to confirm  the mass loading/
opacity relationship.
                                 261

-------
Non-dispersive infrared and thermo-gravimetric analysis of the parti-
culate on a single outlet filter from the run with low fly ash loading in-
dicated that emissions from entrained scrubber reaction products were
in the 0.005  gr/dscf range.  Based on this  single analysis, the outlet
loadings appeared to  be predominately  fly  ash.  Preliminary  results
from air/slurry  entrainment  tests  on  the  TCA  system indicated a
slurry ( ~ 40  percent fly ash and ~ 60 percent slurry reaction products)
entrainment  value between 0.003 and 0.005 gr/dscf. This value is in
agreement with the low reaction product emission value and supports
the conclusion that  over 90 percent of the TCA  emission is  fly ash.
Similar tests are scheduled for venturi/spray tower  system.
The highest emission values experienced  under base case  conditions
occurred with  a  continuous diluted  clarified liquor mist  eliminator
underwash (RunTFG-lC). Changing to an all fresh water mist elimina-
tor -wash scheme (Run TFG-2B) reduced emissions.
Reducing  the circulating  slurry  rate  had no effect  on the  emissions
(Run TFG-2E),  but reducing the  gas rate  (Run TFG-2D) tended to re-
duce the emissions.  MgO addition  (Run  TFG-2A) tended to increase
emissions  although this run did  not have the highest emission rate
since it used a fresh water mist eliminator wash scheme.
Inlet and outlet grain loadings  as a function of  aerodynamic particle
size were  calculated    using  mean values  of the  measurements  for
each run and  are plotted for runs with high fly  ash loading in Figure
7 and for  rune with low fly ash loading in Figure 8.  The inlet distri-
butions were  similar to the venturi/spray tower measurements, con-
firming that the scrubber inlet  size distributions were the same even
though they were taken from the boiler outlet duct at different points.
Mass penetration  as  a  function of aerodynamic particle size is plotted
for the runs with high inlet fly ash loading and low inlet fly ash loading,
in Figures 9 and  10, respectively.  The mass penetration for the high
fly ash loading runs varied from about 40 to 90 percent for 0. 1 micron
particles  to  less  than  2  percent for particles  larger than 5 microns.
For  the low  fly ash  loading runs, penetration was between 25 and 70
percent for  0.1 micron  particles and about 10  percent for particles
larger than 5 microns.
                                  262

-------
For the high fly ash loading runs, emission of particles  less than 2
microns actual diameter averaged  ~ 0.025 gr/dscf  except for the low
gas rate run (TFG-2D) and the  low slurry rate run (TFG-2E) where
the average was higher at ~ 0.037  gr/dscf.  In a TCA, this decreased
removal efficiency would be expected with slurry or gas turndown be-
cause of decreased sphere activity inthebeds. Emission of particulates
with diameters greater than  2 microns averaged ~ 0. 028 gr/dscf except
for the base case  with a  continuous mist eliminator wash  (TFG-2C)
and the MgO  addition run  (TFG-2A) where  the emissions averaged
~ 0.055 and' ~ 0.038 gr/dscf,  respectively.  The low fly ash loading
run averaged 0.003 gr/dscf less than 2 microns actual diameter and
0.025  gr/dscf  greater than 2 microns actual diameter.
From this  data it  would  appear that slurry  entrainment after  eva-
poration in the reheater contributes mostly to  the  large size fraction
of the emissions.   For example, the  continous diluted clarified liquor
under-wash (TFG-2C versus TFG-2B) consisting of water and dissolved
solids   significantly  increased  emissions only  in  the  greater than 2
micron fraction.  Also, when the gas rate (TFG-2C versus TFG-2D)
or the liquor rate  (TFG-2C versus TFG-2E)  was  reduced, the fine
particulate increased but the larger particulate decreased.  This prob-
ably resulted from a decrease in slurry entrainment .  MgO addition
which increased the total dissolved  solids also resulted  in an increase
in emission for the larger  sizes.
                       System Comparison
Run VFG-1I  and TFG-2B from Tables 1 and 2 were chosen to compare
the relative  flue gas cleaning capabilities of the two  systems under
typical  operating  conditions  for  attaining 80  percent SC>2  removal.
These runs  were made under conditions of similar percent solids re-
circulated (15 percent) and mist eliminator wash scheme (fresh water
wash with the bottom wash  off during testing).  The mist eliminator
type (threepass, open-vane chevron) was  the same for both systems.
The  venturi/spray  tower  gas rate was  35,000 acfm compared with
30, 000 acfm for the TCA  system. However,  particulate removal in
                                  2b3

-------
the venturi/  spray tower  system appeared to change  little with  gas
rate (Run VFG-1C versus  Run  VFG-1D) and a  correction  to  30,000
acfm was assumed to be neglible.
Comparison of the two systems showed that the estimated average out-
let mass loading value was ~0. 026 gr/dscf for the venturi/spray tower
system versus  ~0.048  gr/dscf for  the  TCA system.   The average
mass emission of  particulate less than 2 microns actual diameter was
~0. 016 gr/dscf for the venturi/spray tower versus ~0.026 gr/dscf for
the TCA system.  Mass emission greater than 2 micron actual particle
diameter was  ~0. 008 gr/dscf on the venturi/spray tower system ver-
sus ~0. 027 gr/dscf for the TCA system.  The less than and greater
than 2 micron data come from the impactor runs which add up to  slight-
ly different values than the mass loading averages.
It  should  be  emphasized that this comparison is  only valid for the
scrubbers  at Shawnee under the conditions specified  and  is not meant
as a blanket  generalization about the relative capabilities  of the  two
types of scrubbers. Possible design modifications and  economic factors
must also be considered on any final system evaluation.
Conclusions
Analysis of the flue  gas emissions  at the Shawnee Test Facility  has
shown that mass emission is  higher for  the  TCA system  than for the
venturi/spray tower  system under typical conditions. Indeed, the TCA
mass emission approached or exceeded the EPA New Source  Perfor-
mance  Standards  for mass  emission while the  venturi/spray  tower
emission was consistently below the  standard.
Scrubber reaction product emission represented only a small percent-
age  of  the total mass   emission.  Size  distribution run results and
scanning electron microscope photographs indicated that most  of the
scrubber reaction product emission was confined to the larger particle
sizes.
                                 264

-------
Additional measurements of the magnitude and size distribution of
the scrubber reaction product emission are to be made.
Measurement of mass penetration through the scrubbers as a  func-
tion of  particle size indicated that there  was no massive generation
of particulate,  (i.e.,  outlets greater than inlets) and that penetration
became significant in the range of one micron and  smaller where it
varied from  25  to  90 percent. Under comparable conditions,  the re-
lative mass penetration  for  the TCA  system was greater than for the
venturi/spray tower system.
                                 265

-------

0.
C!
U,


g
>


O
6
u,
>

u-

6


s
6
u.

Q
g
U

g




CD
6
u,





<
6
>





u


0^-go^ ^£^™°
°- o^ * * ' °° I j *" "^
£ ~ °° u

-.!
!p!.i*=fj!3S

0
g „ g 0 ^ ., t
°- o^T S^^Sf^g^z

*
"u
o

«; " ~° I J J=


!•§»-•. -agfas
s °~ -' " ^ s J «

c
o o «, o
o-troo ti P u M °
o * -.* ^ £ J f
0 c

S ~ "" z .3 J=
0

o _ o v o
° *f o o £ c i. !_> o
° _ ^ O (7- CO O t > J £
M (J

1 «§° | P -« o
^-^^S0"00^^ £ J £
0


0 0 ,0
0 •«• 0 O -^ p ** _, «l
w

£ « c""*1
ou SI | r S
° JC c . o 2 .c
"" £ a * x S: w
1 £** e g1 g «
« ! 5 j A" « u |
*j">^"^ ^ II'S
5 5 i "3 5 ^-Ew<
m^^CC XJS««O
3feK>i C<222
o
0
„, " »"
0^ ' '
o
*"
0
o
r- °
0
-o
o
v - ^
01 — o
o
•al- m
00
o
o
(-J O
fO O

o
r-"
0
0
o
01 J. o
0
(J. *
•n
o
o
. o
o o
o
-c
•* .

o
— ^
"
a
0 O
(T-
^ O
0
* " "?
°" — o
- "".

0
^ *
ffl o
0
m M
~
C IB
W . "o
» ^
i* S
£ £ ^
m *" *
O^H
sC ? ^
a-
pi -c ^ o ^.  Sr
E -
Hi
•= 'E «
- "o
« 0
b ::
0 §

S « 5
BE 1
1

P
-J

£
1
C ~_
S S
I
be
— C
E 2
a -5
U 0
S u
(S I

c
1
&
1
1
•J



•c
•c
O
S

^i
a
(X
m
H








1 ;
1 1 i-
n C O

£ "° *" o
Q MX «




•
i- « 5 P »
iiili
E S.x $ :







-C i „ "M c
" ' r •» » ™
u ^ ^ 2 w "5
«3SS3E°

5 "S
it °" t
— c o i
S f I j
H - JS u







5
£
E
L)
                                    O
                                   VH


                                    CO
                                    a
                                    ni

                                    OC
                                    o
                                    c S
                                    o w
                                      o

                                      H
                                       S.
                                    ni -rt
                                   o  S


                                    S  S
                                   — j  u
                                   h >
                                    ni

                                   H
                       3 £
266

-------




ro
o



b,
ro
6
tx
H




W

H




Q
ro
6
H




U
^
O
H


O
ro

6
H



ro
6

H




Controlled Variables




0 C C
0 Ul 0 •> O O _
e. M- g 2 I ; S £ 1
o -. -. .J «; 01
1 -s


V
CCA
c> o o m
O m O J3 *j t, ra
. r\J ro — ' £ 4) a] * |Z
0 -H - £ P ,C 0
<*» Sue


OJ
c c
o x o 5
° ij § - - S > u 2
g- jh



c c
S ,-, ° °
O^iO A-*-^iQ
• co r-J "•• £ *V o>
N " £ "5



C g

O . O in QC tn i* . , O
. M ro — ' £ u £ ^
° " "" J f


o c d
o m o ,c o o
O . O in BO ~ *- -
°~~ KllKZ



l^tf !„.
" £3 ^ "" E c fti ^

J U


-
u c 4>
Gas Rate, acfm @ 300°F
Scrubber Gas Velocity, ft/sc
Slurry Rate, gprn
Percent Solids Recirculated
Fly Ash Loading
Alkah
Mist Eliminator Configuratio
Mist Eliminator Wash Schem
MgO Addition
o
o o o

o .
°° 0 0 0
. ^" O
CO . O
0 .
r-
o
o o
r- d co*
* J) •* 0
. t o
m . Tf
0 .

0
in o
O m ro

in o xo
(NJ
. Tf 0
O
0
0 O
— '  , O , (jl ,
"* r- o in °^ o
d
CO
^H CO ^H


O^ -^O *X O CO CT-
^ ' ° ' « '
O 00
d
m

oc o p

<^ -C u-, O — . Q,

^' ro d - 0 ^
fM in
O
O
~H 00 fV)
in o

ro r- 3 ° cr- o-
00 i O i • i
^ r- d co °"" •<*•
O CO
d
CO
m o
^ r-' 3 o ~ £
C3 i O , c> i
•^? ^" o r*-  o
                              S -5 c
                             I-H U
                             K  .
                               U
                                         o

                                         to
                                         to
                                         0)

                                         S
                                         rt
                                         C
                                         O
                                        • iH

                                        "rt
                                         N
 U
 rt

 nJ

U
 en
 rt
O
                                           U
                                           H
n)
H
267

-------







6
>












CO
7
a
U*
£*







6















^



(U
u
tS


5; .
H






^.
J2
*

— •
JC
U
0)
CQ


5 J
H *

^_Q
J
4)
.c
u
QJ
CQ

< .
H









u
a
en

(M




~



3
•i
'







_.




-'



D
ra

r.





0
(4













. . . . 1 . o

m .
O

O •* •— * ^ on ^ in
^* i— i m m o •«*< o
0

f"~ (M vO O^ O fsj on
ro o ^ ro ro ^c o
o




Tf o oo -— ro (M
. . . . ' . o
o — m on ' oo o
o

O -^ r— i *~- ' Is" on on
• . • . . O
O on m — i •— i I-H o
O


... . o
vC o r- «-i in oo o
— (VJ
o

^ **. ~. °! | ^ o
o

r*- on og o 00 in if}
O

^* ^ rO O^ in in oo
r—i rn .
o


u at
p d
Li t] "-1
°^«o«H °'° *° *«



> ^ & ^ — *; o*^fe^ "5^^
« »  a
'^ o
rt ^
.t^ in
c

p o
cr *j
i 3
c ^
in
II n)
(M ^
C u
nj p
I o
t) >r
£ S
S o*
£ ^
4) U
O + *•"*
0\
* O ^
ii ro -^
x o
~ N Ul
-^ DC
"* S
" o" £
3 W tj
4J «] jj
S III «
01 ,", C
"D U 0
0 u
v v n
E E *
o 5 o
£ tn M
H < S

— ' fM


in
(D
O
                                 CuO

                                 .S  a
                                 T3  O
                                 n)  'ri
                                 to  0) •

                                 W  JH
                                    rt

                                    3
                                .H  «J

                                 t-l
                                 "S  -5
                                 0)     o


                                 III
                                 «)  O JH
                                 co  w a
                                 a
                                 t—t
                                 (U
268

-------
)U —




O
T3
O) 10-
— *
O
in
Q
TO
o
T3
5
TJ
-~ 01-
0
z

Q
0
CO
CO
<
_l 001 -
<
1-
z
LU
DC
LLJ
LL
LL
Q

0001 -


1 i i | i i i 1 1 i i i ] i i i 1 1 1 1 1 | i i i i | i i I | I i i (.
P VFG 1A _
O VFG 1C Q
D VFG 1D (ffa
0 VFG IE ^
A VFG IF
V VFG 1G
: O VFG 11 £ 0 ~
*3 VFG IP ^^ WV°
Q ^ A
NOTE SOLID SYMBOLS ARE 0 QQ -
OUTLETS. OTHERS _£_
ARE INLETS 0 0

•
O ^
- V A f^i ^
: o u :
TT
o *»

0
n ^7 ^
A A
Qj §
* * 1
f + % ' * J

• ^
_ . ^
A
^ ^
u • •

^
- —
: T '
-
i i i I

 001
                01             10              10
           AERODYNAMIC PARTICLE DIAMETER, microns
100
Figure 1.   Mean differential mass loading versus aerodynamic
           particle size for all Venturi/Spray Tower system
           high fly ash loading runs.
                           269

-------
      10
 u
 
 o
C3
Z

Q

O
Z
UJ
cc
LLJ
U.
LL
      1 0
     001
    0001 --
   0 0001
        001
                    |
               O VFG  1B (10-22 10/25)

               A VFG  18(10/2610/28)

               NOTE SOLID SYMBOLS ARE OUTLETS

                     OTHERS ARE INLETS
                                                              !   1  1   1 1 1 ..
                              i  i  i
                        01               10              10

                      AERODYNAMIC PARTICLE DIAMETER, microns
                                                                        100
         Figure 2.  Mean differential mass loading versus aerodynamic

                    particle  size for the  Venturi/Spray Tower system

                    low fly ash loading run.
                                    270

-------
    O  Q  ui u_ O  —  a.


    u.  C3  u. u. u.  u.  u.


    o n o < <
-•8
                                                    cm +o  p    o
                                                on
                                     
                                                                          i—t
                                                                          O
n)
a
o  *-<
>H  <1J
0)  £
rt  o

w E-i
2  ,
m  ^
^  rt
0)  ^
>  cx
  to
a •--
                                                                            a
                                                                             bD

            to r -S
            w C T)
            crj rt 
-------

8 8
0 0
(N (D
O 0
i m m -
C3 C3
u. u.

'

o <

o <


o <
* '
B
• o < '•


o <


- O < :
•
«
0 <

o <

- o ^
.
: o< :




-8
' t~

w>
c
o
u
1
DC
UJ
|_
. 0 UJ
'. "~ >E
^
; 5
UJ
1
O
h-
oc
Q.
o
-2 |
z

Q
O
DC
UJ
^

o

• o

o

u
• ^H
a
0
•i-i
a
(TJ
cj
^»
"o
Q)
rt
w
w
t,
a)
>
G
O
'•^
^
•4->
(U
C
ex
c
(D
u
h
a
rt
s
-*
(1)
tn
be
.1-1


tuO
C
• rH
ni
^H
(ti
^N
^^
^H
£
.2
h
J
o
H

ni
^
'Jn
3
C
.>
S2
•*-!
LJ r ^

6 c
.2 5
£ rt





'NOIlVai3N3d SSVIAI
             272

-------
Figure 5.  Particulate on third stage of outlet impactor during
           Venturi/Spray Tower  run VFG-1A. Magnification 2000X.
Figure 6.  Particulate on  sixth stage of outlet impactor during
           Venturi/Spray  Tower run VFG-1A.  Magnification 5000X.
                             273

-------
     10
     10
o
in
Q
o>
^

O
^i   01  --
Q

O


8
Z
ill
cc
LU
LL.
LL

Q
    001
    0001  --
   00001
             NOTE SOLID SYMBOLS ARE
                 OUTLETS. OTHERS ARE
                 INLETS
       001


                       01               10              10

                    AERODYNAMIC PARTICLE DIAMETER, microns
                                                                     100
        Figure 7.  Mean differential mass loading versus aerodynamic

                    particle size for all TCA system high fly ash load-

                    ing runs.
                                   274

-------
      10
     10 - -
g
Q
O)
o
o
5
o
     01 --
    0 01
LU
DC
LU
LL.
LL

5
    0 001
   00001
              NOTE SOLID SYMBOLS ARE
                  OUTLETS. OTHERS ARE
                  INLETS
                                                              to
                                               D
                                     D

                                      A
                      a
       001
                       01               10              10

                    AERODYNAMIC PARTICLE DIAMETER, microns
                                                                      100
        Figure  8.  Mean differential mass loading versus aerodynamic
                    particle size for all TCA system low fly ash load-
                    ing  runs.
                                       275

-------
   <  DD  O  Q  LU  U.
   CVJ  CM  CM  CM  CM  CM

   CD  CD  CD  CD  O  CD
   U.  LL.  LL  U.  LL  LL
   I-  t-  I-  H-  I-  I-



   k  o  n  < o  >
                                                                   --S
                                       n   t*xa  o


                                    Bfc-
                                
                                                                                f— 1
                                                                                 u
 rt
 CX
 u
• i-l
 a
 rt
 C
                                                                                   be
 en
 h
 1)
                                                                                ni
                                                                                ^
 C ^
 0) ^
 u T!
en
to  0)
r«  N

^  to
                                                                                bo
                        'NOIlVai3N3d SSVIAI
                                         276

-------
                                       -.8
              D<  O  O



               rO
       ? ni
       to o
                                                  C
                                                  0
                                                  a

                                                  u
       en
       co D

       rt N
      ^H •!-<

      A 03
       0)

       ^



       CuO
'NOIlVai3N3d SSVIAI
                277

-------
References









1.  Bechtel Corporation,  EPA Alkali Scrubbing Test  Facility:  Sum-




    mary of Testing  through October  1974,  EPA Report 650/2-75-047,




    June 1975.




2.  Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced




    ProgramFirst Progress Report,  EPA Report  600/2-75-050,  Sep-




    tember 1975.




3.  Bechtel Corporation, EPA Alkali Scrubbing Test Facility; Advanced




    Program Second Progress Report,  EPA  600/7-76-008,  September




    1976.




4.  M. P. Schrag  and  A. K. Rao,  Fine Particle Emissions Information




    System; Summary Report (Summer 1976), EPA Report 600/2-76-1 74,




    June 1976.




5.  Federal Register, Vol. 36, No. 247, 12/23/71, p.  24888.




6.  TRW Corporation, Procedures for Aerosol  Sizing and SO^ Vapor




    Measurement at  TVA Shawnee Test Facility, Document No.  24916-




    60390-RU-01,  1977.




7.  D.  B. Harris, "Tentative procedures for particulate sizing in pro-




    cess streams - cascade impactors", IERL,  EPA , Research Tri-




    angle Park, N. C.,  ppm. 30-32, February,  1976.




8.  R. H. Perry, Chemical Engineers' Handbook, fifth edition, McGraw-



    Hill Book Co. , New York, 1973, p. 5-61.
                                  278

-------
                   THE WATERLOO SCRUBBER
                      Donald R. Spink
                   Francis A.L. Dullien
                     Peter L. Douglas
            Department of Chemical Engineering
                  University of Waterloo
                  Ontario, Canada N2L 3G1
                         ABSTRACT

     The body of information presented in this paper is
directed to all those poor souls who are faced with tough
particulate removal problems and frustrated because only
expensive and high energy-consuming solutions are available
to contend with the problem.  A simple and yet highly
effective scrubber is described with examples detailed for
application to treatment of particulate-laden hot gases
and implied for treatment of cooler gases, e.g.,  those less
than 100°C.
                            279

-------
                        THE WATERLOO  SCRUBBER
Introduction
          The purpose of this paper is to

          a)  describe the Waterloo Scrubber and  point  out  its many  features

          b)  present laboratory and plant data to  illustrate  the  abilities
              of the Waterloo Scrubber to effectively remove fine  particulate

          c)  present a theoretical, mathematical model which  appears  to
              account for the removal efficiencies  observed

Figure 1 is a photograph of our laboratory scrubber installation which
clearly illustrates the simplicity of this scrubber.  The main components
are the duct work, the slightly modified fan and  a  means to introduce  a fine
mist or fog into the duct upstream from the fan.  The cross-section  of the
Waterloo Scrubber, shown in Figure 2, illustrates its simplicity.  The  major
modification to the fan is to include a drain at  the  lowest point  in the  fan
casing as shown in Figure 2, along with the water seal  used to receive  the
agglomerated spray and contained particulates.  A source for introducing  a
fine spray of water or other liquid is also shown.  While we have  studied a
number of fine-spray-producing devices, the so called "sonic"  nozzles  appear
to be most suited for two reasons:  their energy  consumption is modest, and
wear, tear and maintenance is minimal.  One disadvantage of sonic  nozzles
is that, when properly tuned, they can be somewhat  noisy and therefore
distracting.  Insulation of the duct may be required  depending on  the  loca-
tion of the scrubber.

          There are three criteria which are necessary  for  the proper
operation of the Waterloo Scrubber:

1)  The mist or fog of droplets introduced must be  fairly fine, i.e.,
    normally on the order of from 5 to 50 ym in a reasonably uniform
    distribution.

2)  The fan must turn at a relatively slow speed  so as  to provide  not  only
    the necessary degree of turbulence but of even  greater  importance,  the
    requisite residence time for the contact, agglomeration and elimination
    of the liquid mist and particulates originally  contained in the  dirty
    gas stream.

3)  The number of liquid mist droplets introduced to  the system must be in
    proportion to the number of particulates to be  removed  from the  system;
    however, as dust loadings increase the ratio  of liquid  droplets  to  dust
    particles can be reduced and visa versa.

          If the mist or fog of liquid is introduced  into the  system by
condensation of water vapour, such as would occur if  a  hot, saturated  gas
stream were suddenly cooled by the introduction of  a  small  stream  of cold

                                    280

-------
air, in that the vapour is caused to condense in situ, such condensation
will normally take place on the surfaces Of the dust particles within the
duct, thereby causing these to grow.  This is especially true when the
dust particles are hydrophilic and even more so when they are hygroscopic.
Under such conditions, less energy is consumed and it is found that the
amount of water so produced on condensation need not be as great an amount
as would have to be furnished by an atomizing nozzle.  We'll  come back to
this point shortly.

          If a finely dispersed liquid mist is introduced into most scrubbers,
a significant number of the fine mist "particles" will not be captured. Thus,
the removal of acid mists, for example, is often a serious problem to many
chemical and metallurgical industries.  Conventional "demisters" are also
ineffective in the removal of fine mists or fogs.  Thus, one  of the
important features of the Waterloo Scrubber is its ability to remove fine
mists or fogs from a gas stream.


Experimental Methods


          Standard testing procedures were employed for all experimental
studies reported here.  These procedures include isokinetic sampling and the
use of Joy WP-50 particulate train.  Calculations were based  on those out-
lined by the Joy Flanufacturing Company in their bulletin WP-50.

          All data was statistically examined to ascertain its reproducibility
and level of confidence.

          Only a limited number of particle size distribution experiments
have been conducted, primarily due to the tedious nature of.such work and
the high degree of uncertainty associated with the results^ '.  Prior to
late 1976, this type of data was determined the "hard way".  A dispersed
sample was first collected on a polished stub (a source of much controversy),
gold plated (vapour deposition), subjected to a scanning electron microscooe
photograph which was then analyzed on a Quantimet.  Some confidence was
initially established with this approach when results matched the manufac-
turer's particle size distribution data very closely for several different
pigments tested.

          Late in 1976 an Anderson 8 stage, cascade impact sampler was
obtained.  Initial use of this equipment has been primarily in studies to
improve the methods employed to introduce and disperse the test dust into
the system.  A few scrubber outlet samples have also been collected on the
Anderson sampler to determine individual particle size removal efficiencies.


Results
          The method of introducing test dust into  a scrubber system is  one
of the most critical steps in the study and development of particulate

                                    281

-------
removal systems in the laboratory.  We have employed numerous  methods  in
attempts to develop a superior system and by continuous improvements have
now achieved the requisite excellent dispersion of fine powders by the
equipment shown in Figure 3.  Particle distribution data using an Anderson
Cascade impact sampler have confirmed the excellent dispersion achieved
with the simple equipment shown in Figure 3.  The vibrating feeder delivers
a uniform flow of the dust to the throat of an air venturi  where a high
velocity air jet draws the dust into the system while at the same time
effectively dispersing it in the duct of the system.

          The test dust employed to the greatest extent in  our studies is
a pigment-grade aluminum silicate; Figure 4 shows the characteristics  of
this dust.  Also shown on Figure 4 is a typical particle size  distribution
(count not normalized) for an outlet sample from the Waterloo  Scrubber.
For this particular experiment and sample, all particles greater than  1.2 pm
have been completely removed; this phenomena has been observed for many
years with the Waterloo Scrubber.

          The ability of the Waterloo Scrubber to completely eliminate a
fine mist or fog is believed to be responsible also for the removal of fine
particulates in the gas stream.  (At the meeting, a short film will be shown
at this point which illustrates the ability of the Waterloo Scrubber to
completely eliminate a fine mist or fog from a gas stream).

          The observed instantaneous elimination of a fine  mist or fog by
the action of a slowly-turning fan has been analyzed in terms  of several
"classical" mechanisms of coagulation and precipitation of  aerosols, but
none of these classical mechanisms account for more than a  fraction of the
effect observed.  The elimination of the fine mist can be fully explained,
however, if it is assumed that turbulent diffusion on a scale  comparable
to the effective stopping distance of the larger drops present in the  mist
is responsible for the fast agglomeration of the mist to produce larger
droplets which are easily removed by the centrifugal action in the fan.

          Physical justification for this assumption may lie with the  postu-
lated mechanism whereby, for very short periods of time, large directional
velocity differences exist between droplets of different size  such that the
simultaneously existing turbulent diffusion of the particles is made far
more effective than has been assumed by existing theories,  resulting in a
much higher rate of coagulation (or agglomeration) than would  be observed
either due to a similar velocity difference between the different sized
droplets in the absence of turbulence or as a result of turbulent diffusion
in the absence of a large velocity difference between the droplets.

          The resulting mathematical model, written as a symmetrical expres-
sion for the rate of disappearance of small and large droplets (particles)
is as f o 11 ows (5 ) :


                                DL + °S , £ .1/2 (XL - xS}eff     n    .2  m
                       dt --- * - 2 - W    - 12 - nL n$ n  t   (1)
                                                     eff
                                   282

-------
where n = number of drops in a unit volume
      t = time
      D = diameter of drops
      e = energy dissipated per unit mass of medium
      v = kinematic viscosity
      x = distance travelled by particles (stopping distance)
      n = capture coefficient

subscripts

      L = large
      S = small
    eff = effective

This model is based on the assumption that only two sizes of droplets are
initially present and in equal number.  The introduction of a true size
distribution f(0)dD of the drops initially present so as to obtain the
droplet size concentration at any time t is presently being developed.

          Numerical integration of equation (1) over very short time periods
to obtain an approximate solution gives new drop sizes D' and Di and cor-
responding new concentrations n/ and n^; this process can then Be repeated
for successive short time perioas to develop the ultimate particle size
distribution at any time.  Such calculations employing time increments of
0.0005 seconds indicate that virtually all of the particles are large
enough after relatively short residence short residence times to be
removed by centrifugal action in the fan.  While such an approach provides
only an approximate solution, the results relate very closely to that
observed.  Further development and testing of the model are now in process.

          We have heretofore emphasized the necessity of a slowly turning
fan.  Application of equation (1) is in agreement with this concept where
t ,;,:, the effective residence time of particles in the fan, must be suffi-
cient to allow the interactions theorized to take place.  Such results are
expressed in a different fashion as shown in Figure 5.  The relationship
shown should not be considered as exact but "illustrates" in pictorial    ,r\
fashion what we have observed and what has been reported in the literature1  '.
At the higher speeds, turbulence increases at the expense of residence time.
It should be pointed out that even when the fan is turning relatively slowly,
a highly turbulent condition still exists within the casing of the fan.   As
an example, a fan turning at 600 rpm and delivering 800 cfm of air will  have
a fan Reynolds Number around 105; a larger fan turning at around 230 rpm
and delivering 50,000 cfm of air will  have a fan Reynolds Number in excess
of 10 .   The effective residence time for both situations will be approximately
the same.

          Another relationship which is important is the very small amount
of wateK normally employed in the Waterloo Scrubber.  Injection of less  than
0.03 dm  of water per m^ of gas (approximately 3 Ibs. per 1000 ft3) in the
form of a mist consisting of droplets  ranging in size from about 5 um to
about 65 ym and with a number average  droplet diameter of about 20 to 25 ym


                                   283

-------
is normally effective for the removal  of fine particulate.   Thus,  in  the
process of eliminating participates from a gas stream,  a very  small  and
thus relatively easily handled aqueous discharge stream results.

          All of our earlier work employed Spraying  Systems  pneumatic
atomizing nozzles.  Figure 6 illustrates the relationship between  average
mist droplet size and air pressure used for the Spraying Systems nozzle.
Figure 6 is related to Figure 7 which  shows the number  of mist droplets
produced per minute, also as a function of air pressure employed when
spraying about two pounds of water per minute.  Note that at normal
operating conditions, some 100 billion water droplets  are produced per
minute.  The drop size distribution obtained with the  pneumatic nozzle
(Spraying System! at,an air pressure of 485 kPa and  at  a water flow  rate
of 1.1 x 1Q~2 dm  s"  is shown in Figure 8.  The droplet diameters were
measured by using the magnesium oxide  method developed  by May''),  whereby
the airborne droplets are allowed to strike a layer  of  magnesium oxide
on a glass slide.  The resulting permanent impressions  were  then measured
under a microscope.  Work is currently being conducted  in obtaining  similar
information for various sizes of sonic nozzles when  operated under different
conditions.  It appears that equivalent or better performance  can  be  expected
when using one or more of the smaller sonic nozzles  with less  than 101 kPa
air pressure.  The larger sonic nozzles don't appear to be as  effective  at
producing very fine mists as the smaller sonic nozzles  and normally  require
higher air pressure.  Thus, in some cases, several  of  the smaller  sonic
nozzles might be more affective than one larger nozzle.

          The effect of dust loadings  in the dirty gas  stream  and  the
effect of air pressure with a Spraying Systems nozzle  on particulate  removal
efficiency is shown in Figure 9 (using aluminum silicate pigment).  To
achieve higher removal efficiencies when dust loadings  are very small,
relatively speaking, many more fine water droplets would be  required. When
dust loadings are very high, removal is made much easier due to particle-
particle interactions which tend to form agglomerates without  requiring
water to assist to any great extent.

          Several additional test dusts as supplied  by  clients  have  been
evaluated in our laboratory facilities.  Unfortunately, we did not charac-
terize these test dusts.  Tests on carbon black and  aluminum oxide,  for
example, each resulted in complete removal.

          A number of tests have also  been conducted on the  gases  eminating
from the "recovery" boiler at a kraft  pulp mill.  In these tests,  which
were conducted at the Consolidated Bathurst mill at  Portage  du Fort,
Quebec, the gases were first treated by an electrostatic precipitator
,-ated at 95% efficiency.  Removal efficiencies of 96%  of particulates
remaining after the electrostatic precipitator were  achieved with  the
Waterloo Scrubber on these preliminary tests.  Unfortunately,  we have not
yet been able to characterize the particulate emission  from  the recovery
boiler of a kraft mill after passing through an electrostatic  precipitator;
however, the indicated mean particle diameter is about

                                   284

-------
0.4 ym.     During tests conducted at Consolidated Bathurst, the gases from
the electrostatic precipitator were received at 138°C (280°F).  We found
that when a fine water spray is introduced into a turbulent gas stream at
138°C or thereabouts, evaporation of the smaller droplets takes place in a
small fraction of a second^.  Therefore, before we can effectively
introduce a fine spray into a hot gas stream, we must cool  and humidify the
gases to near their dew point.  This can be accomplished in a number of
ways.  In the initial kraft mill tests, we accomplished this task by
inserting a total of four Spraying Systems atomizing nozzles in the duct
carrying the hot gases; we found that the hot gas was at its dew point
after three nozzles so the fourth was placed several meters from the fan
scrubber.  Unfortunately, this method of gas cooling is fairly high in its
consumption of energy since each Spraying Systems nozzle requires about
520 kPa air to be effective; the calculated energy to compress the air
used in each of the nozzles is about 370 W (1/2 HP) depending on the
actual amount of air used and method of compression (we did not measure
air consumption for these tests but used manufacturer's information).
It could perhaps be argued that the cooling and humidification of the gases
is unrelated to the actual particulate removal step, but nevertheless, it
is a required step.  Total power consumed for the initial test at Consolida-
ted Bathurst probably falls between 3.95-5.53 kwh/m3 (2.5 and 3.5 HP/
1000 ft^/min) of gas scrubbed vs. about 1.5 kwh per m^/m\n  (approximately
1 HP/1000 cfm) when cooler air is being scrubbed in the Waterloo Scrubber.

          Once the effectiveness of the Waterloo Scrubber to remove sub-
stantial amounts of salt cake fly ash from the recovery boiler off-gas
was established, our thoughts were then directed to methods to improve
removal efficiencies and reduce energy consumption.  It was felt that
humidification of the hot gases could be accomplished in a  simple packed
tower; energy related to such a device would be that required to lift
water from the sump to a distributor above the packing, a distance of less
than 10 feet, plus energy related to the increase in pressure drop across
the system.  Rough calculations indicate that a total  of less than
1.5 kwh/m  would be associated with the operation of the humidification
tower.

          Once the gases have been cooled and humidified close to their
dew point, further cooling of the gas will produce a condensation effect
thereby resulting in the formation of a cloud of water droplets within
the duct.   Such condensation will preferably take place on  the surface of
dust particles in the gas stream which could conceivably lead to improved
removal efficiencies.  This approach was first studied with the laboratory
scrubber.   The problem of producing up to 1/2 m /s of gas at its dew point
and at approximately 160°F proved to be an interesting problem in itself.
A salamander (gas heater) rated at 210,000 kJ/hr was the initial source
of hot gas; this was cooled and humidified to some extent by the fine spray
from a "sonic" nozzle,  and followed by live steam provided  at 165°C and
about 500 kPa from a 1/2" open line.   With minor adjustments, the requisite
conditions were achieved.  Test dust was introduced as  described (vide supra)

                                   285

-------
In the laboratory tests, the cooling air was allowed to be sucked into the
duct through variable-sized slots.   The slot openings were varied by sliding
a curved cover over a rectangular opening in the round duct.   The formation
of a mist or fog at the point of injection of the cooler air  was  readily
apparent.  Whereas in previous experiments with the aluminum  silicate
pigment as the test dust, efficiencies of about 96% were about the best
achieved without excessive use of atomizing sprays (see Figure 9), all of
the experiments employing the condensation approach to introduction of the
mist within the duct showed at least 98% particulate removal  and  approximate-
ly half of such tests resulted in complete removal of the test dust.

         Accordingly, the installation at Consolidated Bathurst has now
been revised to eliminate all of the atomizing sprays.  Instead,  the energy
of the hot gases will be used to provide the saturated situation, using a
small packed tower for the humidification.  Preliminary tests indicate that
the hot gases can be cooled to their dew point with less than 12  &m~'•,
circulating flow when the flow of hot gases was in excess of  1/2  m3s"
(> 1000 ftVmin).

         A schematic flow diagram of the revised installation at  Consolida-
ted Bathurst is shown in Figure 10.  In addition to the benefits  of this
approach as suggested above, water mist, produced by cooling  the  saturated
hot gas stream, will be removed along with particulates by the scrubber
and the combined stream returned by gravity to the sump tank  below  the
humidification tower.  Preliminary calculations indicate that slightly
more water is evaporated in the humidification tower than is  removed by
the cooling-scrubbing process so that a small amount of water must be
added to the system.  This means that the strength of the Na^SO,  solution
contained in the sump tank will slowly build up at a rate proportional to
the amount of Na?SO. removed from the hot gas stream.  Na?SO. exhibits a
solubility of 44 kg/100 kg FLO at 100°C, so we can assume that the con-
centration of Na2S04 that can be recirculated to the humidification tower
without problems would probably approach 40 kg per 100 kg water.   At
steady state conditions, a constant and fairly concentrated quantity of
iMapSO. solution will be removed from the sump tank and this volume must
also Be replaced by an equivalent amount of water to maintain a proper water
balance.  The NaSO. solution will be recycled back to the mill where
further concentration may be conducted before addition to the strong black
liquor which is fed to the recovery furnace.

         To measure the energy efficiency of the Waterloo Scrubber relative
to other scrubbers, the parti cul ate removal efficiency must first be
converted to a dimensionless number referred to as a 'transfer unit1.  The
transfer unit, N. , is expressed as
where n is the collection efficiency and 1  - n is the penetration.
                                   286

-------
An efficiency of 96% corresponds then to 3.22 transfer units,  98% to 3.91
transfer units and 99% to 4.61 transfer units.

         For a number of different scrubbers  operating under different
industrial conditions, Semrau has observed that
for which 'P.' represents the theoretical  power input to the  scrubbing
process, and  'a'  and 'b' are constants related to properties  of the par-
ticulates being scrubbed^'.   Figure 11  shows a plot of this relationship
between efficiency and horsepower/1000 ft^/min for wet scrubbers  when  used
to scrub the particulates from a pulp mill  recovery boiler emission.   The
initial tests at Consolidated Bathurst (96% removal efficiency) consumed
between 2.5 and 3.5 HP/1000 cfm and this  information is shown on  Figure  11.
Also included is  the projected point for  the revised system at Consolidated
Bathurst.  Whether Figure 11 would apply  to the capture of particulates
after passing through an electrostatic precipitator is not known; certainly,
particle size and concentration are greatly different.

         To date, all testing of theJdaterloo Scrubber has taken  place over
a rather limited  range of 0.25-1.0 m s"  (500 to 2000 cfm) so no  real
evaluation of scale-up factors has been attempted to date.  Based on the
mathematical model and numerous other studies, constant residence time is
felt to be the most critical element in scaling-up.  Interestingly enough,
as larger gas volumes are required to be  handled, larger fans must be  used
at lower speeds;  however the turbulence in  the fan, as expressed  by the  fan
Reynolds Number,  increases (vidi  supra).   Figure 12 illustrates the rela-
tionship between  fan speed and fan capacity for Waterloo Scrubbers having
constant residence time.  The relationships described above might indicate
a potential  for improved particulate removal  efficiencies with scale-up  to
handle larger volumes of gas.


Conclusions
         The following items summarize the  advantages  inferred  or confirmed,
for the Waterloo Scrubber over competing devices  for fine  particulate
removal:

         1)   low energy consumption
         2)   low capital  cost
         3)   low operating cost
         4)   low maintenance
         5)   low water usage
         6)   highly effective for the  removal  of  sub-micron  particulates
                                    287

-------
While many are complaining about the high cost of meeting environmental
requirements, the Waterloo Scrubber appears to be a very simple yet highly
effective solution to many air pollution problems.

                            Acknowledgments


         The authors are grateful to the Committee on Pollution Abatement
Research of Environment Canada for their support of the research and pulp
mill testing programs described and to the fine people in the Consolidated
Bathurst kraft mill at Portage du Fort, Quebec, who have co-operated with
us in such an admirable manner during the test program.  The authors also
acknowledge the invaluable assistance of the many co-op students at the
University of Waterloo who worked on this project and in particular, to
Gordon Berthin for his dedicated efforts.

-------
4J
 ta
 o
_a
 ns
 s-

 O)
-Q
J3


 S-
 O

CO


 o
 o
 at
-i->
 to
 ai


 3
 CD

-------
M
l"~" 1
p.—
^v

V
•#
\i
 \  I
                Ul

                N

                8
                O
                z
                N
                                 S-
                                 Ol
S_
o
00

o
o

S-

(U
T3
                                 (J
                                 (1)
                                 in
                                 2
                                 c_>
  290

-------
\
                                                                                 C
                                                                                 O)
                                                                                 3
                                                                                 cr
                                                                                 CD

                                                                                 en
                                                                                 c
                                                                                 OJ
                                                                                 3

                                                                                Q
                                                                                n


                                                                                2
                                                                                3
                                                                                O)
                                     291

-------
                                                                  0)

                                                                  c
                                                                 -a
                                                                  ai
                                                                  N
                                                                     o
                                                                    oo
                                                                  O  O
                                                                  c  o
                                                                  (/i  O)
                                                                  C 4->
                                                                  o    C
                                                                 CO  O

                                                                 "O  01
                                                                     OJ
                                                                 —
                                                                 N  Q.

                                                                 01  (O
                                                                     OO
                                                                 01

                                                                 "u  o
                                                                 •r-  3
                                                                 •!-> TD
                                                                 i-
                                                                 (O +J
                                                                 Q-  OJ

                                                                 i— 4->
                                                                 (O  3
                                                                 0  O

                                                                 Q--0

                                                                 I—  
-------

t
>_
0
Z
UJ

0
u_
u_
Ul
^



100
90

80


70

60
50

WATERLOO SCRUBBER
\ + H2S04 AEROSOL l-8/im
\ 0 NH4oe O.I-2/zm
\
\


^ *--•
\ ^^ "~^
^^^.^"^
^^^-~~~Qr~~~"~'
I i i
300 600 1200 1800



	 	 . — **^
.^-^
^^^^


	 — —

i i
2400 3000
                              RPM
Figure  5.    Fine participate removal  efficiency as related to fan speed.
                                   293

-------
   180
   160
§
k_
o

1
o:
LU
»-
UJ
O
   120
   100
LJ  80
CL
O
cr
o
 o
Q
   60
   40
   20
     0
 Figure 6.
                 PREDICTED DIAMETER

                 (SAUTER MEAN DIA.)


                 MEASURED  DIAMETER

                (SAUTER  MEAN DIA.)
   20     40     60     80

   P,  AIR PRESSURE ,   ( psig )
100
120
Water droplet size vs. air pressure for Spraying Systems air

atomizing nozzle.

                294

-------
   10000
o>

 'O
    loo-o
 UJ

 o_
 O
 o:
 o
     10.0
 CD
      1-0
       0.1
         0
      20      40       60       80

       P,  AIR  PRESSURE  (psig)
100
120
 Figure 7.
Number of water droplets produced per minute with the Spraying
Systems air atomizing nozzle.
                               295

-------
                                                     m
                                                     to
                                                     o
                                                     10
 m
 m
                                                     o
                                                     m
                                                             1C

                                                             Q.
                                                             CO
                                                     o
                                                      N
         QJ O
         i— C
         a.
         O O)
         S- C
         T3 -r-
          N
         CT)-r-
         C E
         •r- O
         S •!->
         o 
         +j +J
         in i/>
         •r- >-)
         n: oo
                                                     m
                                                             00
O
CM
  U)        CM         CO         sf


3QI1S  NO  S131dOdQ  JO 'ON
o
                         296

-------
            'AON3IOIJd3
            o
            10
  o
  ro
8
                                          o
                                          O)
                                          M


                                          O


                                          II
O
6
o
                                                                       3
                                                                       T3
                                                                   ro    °
                                                                 PC   J
                                                                 o \   o
                                               CD


                                               Q


                                            9§

                                            ""  r-
                                               CO
                                               z>
                                               Q

                                               t-
                                               LJ
                                                                       >*- i
                                                                        (O ID
                                                                         t.
                                                                        >) CL
                                                                        U U1
                                                                        O 4->
                                                                        •r- 03
                                                                        0) 
                                                                          C
                                                                        r— • 'r—
                                                                        10 M-

                                                                        O t-
                                                                          O
                                                                        2.

                                                                        01
                                                                        r— ID
                                                                        3
                                                                        O"O

                                                                        £S

                                                                        10 O)
                                                                        a. c
                                                                        CTl

                                                                        s
                                                                        3
                                                                        cn
IO
m
oo
   o
   oo
                                                               in
                                                               CO
                             lVAOW3d

                                  297

-------
                                                            00

                                                            •a
                                                             c
                                                             o
                                                            o
                                                             O)
                                                             c
                                                             s-  c
                                                            TJ  O

                                                             U  +->
                                                            •t—  to
                                                            +->  r—
                                                             IO  r—
                                                             E  <0
                                                             O)  +J
                                                            JC  10
                                                             O  C
                                                             2
298

-------
   10
~ 8
5E 4
z>

a:
LJ
u_
(A)
UL  2
O
Ld
CD
o PRESENT  OPERATING

  CONDITION

A PROJECTED OPERATING

  CONDITION
                                               I
                     2               4         6      8    10

            TOTAL  POWER  CONSUMPTION (hp/IOOO  cfm )



        Figure 11.  Semrau plot for the recovery boiler on a kraft pulp mill.


                               299

-------

800
Q
u} 600
Q_
"400
Z
jj 200

FAN S\ZE

AW 30
•» AW50
^X^^ AW 70
^X^ AW 120
^^^^^__
i i i
RPM

600
420
330
260

~
cfm

750
3700
9500
48000


      1000   10,000     100,000
      FAN  CAPACITY  - cfm
Figure 12.   Relationship between fan speed and capacity at constant
           residence time.
                              300

-------
                               References
1.    F.A.L.  Dullien and D.R.  Spink,  U.S.  Patent Application, Serial No.
      354,638, Gas Scrubber.

2.    F.A.L.  Dullien and D.R.  Spink,  U.S.  Patent Application, Serial No.
      732,383, Gas Scrubber.

3.    P.L.  Douglas, F.A.L.  Dullien  and  D.R.  Spink,  "An  investigation of the
      operating parameters  of  a low energy wet  scrubber for fine particulates",
      Can.  J.  Chem. Eng., 54,  173 (1976).

4.    T.S.  Munro, "Construction and preliminary investigations of a wet
      scrubber of low energy requirement", Masters  Thesis, Department of
      Chemical Engineering, University  of  Waterloo  (1972).

5.    F.A.L.  Dullien, "The  rate played  by  eddy  diffusion  in aerodynamic
      capture  of particles", Particle Technology, Nuremberg, European
      Congress:  "Transfer  Processes  in Particle Systems" Heat, Mass and
      Momentum Transfer, March 28-30, 1977.

6.    G.H.  Petersen, U.S. Patent 3,653,187 (April 4,  1972) "Apparatus for
      agglomerating suspended  matter  out of  gases and vapours and/or for
      absorbing gas components".

7.    R.K.  May, "The measurement of airborne droplets by  the megnesium oxide
      method", J. Sci.  Instr., 27,  128  (1950).
                 _._    ,     ,.__.-   -"t^-

8.    C.W.  Lear, W.F. Krieve and E. Cohen, "Charged droplet scrubbing for
      fine  particulate  control", Symposium on Electrostatic Precipitators
      for the  Control of Fine  Particulate, EPA-650/2-75-016 (1975).

9.    K.T.  Semrau, "Correlation of  dust scrubber efficiency", J. of Air
      Poll. Contr. Assoc. ,  10, #3,  200  (1960).
                                      301

-------
              FINE  PARTICLE  CONTROL WITH UW ELECTROSTATIC SCRUBBER
                                  M. J.  Pilat
                                 G. A. Raemhild
                            University of Washington
                               Seattle, Washington
                                  D.  L. Harmon
                     U.S.  Environmental  Protection Agency
                     Research  Triangle Park, North Carolina
                                   ABSTRACT
The UW Electrostatic Scrubber system has been installed in two field portable
pilot plants, each in a 40 foot trailer.  The UW Electrostatic Scrubber
involves the use of electrostatically charged rater droplets to collect air
pollutant particles charged to a polarity opposite from the droplets.  The
Mark IP UW Electrostatic Scrubber pilot plant has been tested at a coal-fired
boiler, a magnesium sulfite recovery boiler (pulp mill), and a hog fuel
(wood waste) boiler.  The Mark 2P UW Electrostatic Scrubber pilot plant has
been tested on dioctyl phthalate aerosol, on the emissions from a pulverized
coal-fired boiler, and on the emissions from an electric arc steel furnace.
Measured overall particle collection efficiencies range from about 25 to 99.8%
depending upon the electrostatic scrubber operating conditions and the inlet
particle size distribution.  The test results illustrate that the addition of
electLostatic charging of the aerosol particles and/or the spray liquor droplets
can substantially enhance the collection efficiency for fine particles by wet
scrubbers.
                                        303

-------
          FINE PARTICLE CONTROL WITH UW ELECTROSTATIC SCRUBBER


I.  Introduction

    A.   Objectives of Research Project

        The  objectives of  this on-going research  project  are to  demonstrate
        the  effectiveness  of the UW Electrostatic Scrubber for controlling the
        emissions of fine  particulates, to use the portable 1,000 acfm pilot
        plant in a 40 ft.  trailer to obtain the data needed to design larger
        electrostatic scrubber systems, and to perform preliminary design and
        economic analyses  of full-scale electrostatic scrubber systems.

    B.   Review of Previous Work

        Penney (1944) patented an electrified liquid spray test  precipitator
        involving particle charging by corona discharge and droplet charging
        by either ion impaction or induction.  Penney's system consisted of
        a spray scrubber with electrostatically charged water droplets col-
        lecting aerosol particles charged  to the  opposite polarity.  Kraemer
        and  Johnstone (1955) reported theoretically calculated single droplet
        (50  micron diameter droplet charged negatively to 5,000  volts) col-
        lection efficiencies of 332,000% for 0.05 micron  diameter particles
        (4 electron unit positive charges  per particle).   Pilat, Jaasund, and
        Sparks (1974) reported on theoretical calculation results and labora-
        tory tests with an electrostatic spray scrubber apparatus.  Pilat (1975)
        reported on field  testing during 1973-1974 with a 1,000  acfm UW Electro-
        static Scrubber (Mark IP model) funded by the Northwest  Pulp and Paper
        Association.  Pilat and Meyer (1976) reported on  the design and testing
        of a newer 1,000 acfm UW Electrostatic Scrubber (Mark 2P model)  port-
        able pilot plant.   The UW Electrostatic Scrubber  (Patent Pending) has
        been licensed to the Pollution Control Systems Corporation (of Renton
        and  Seattle, Washington)  for production and sales.

    C.   UW Electrostatic Scrubber

        The  UW (Pilat) Electrostatic Scrubber involves the use of electro-
        statically charged water droplets  to collect air  pollutant particles
        electrostatically  charged to a polarity opposite  from the droplets.
        A schematic illustration of the UW Electrostatic  Scrubber system  is
        presented in Figure 1.  The particles are electrostatically charged
        (negative polarity) in the corona  section.  From  the corona section
        the  gases and charged particles flow into a scrubber chamber into which
        electrostatically  charged water droplets  (positive polarity) are sprayed.
        The  gases and some entrained water droplets flow  out of  the spray chamber
        into a mist eliminator consisting  of a positively charged corona sec-
        tion in which the  positively charged water droplets are  removed from
        the  gaseous stream.
                                     304

-------
GAS INLET
                                                                               GAS OUTLET
              CORONA
          (PARTICLE CHARGING)
 CHARGED  WATER  SPRAYS
(COLLECTION OF CHARGED  PARTICLES
 BY OPPOSITELY CHARGED  WATER  DROPLETS)
MIST ELIMINATOR
                          Figure 1.  UW Electrostatic Scrubber



       II.  Experimental Method

           A.  UW Electrostatic Scrubber Pilot Plant

               The general layout of the UW Electrostatic Scrubber pilot plant (Mark
               2P model) is shown in Figure 2.  The system (in the direction of gas
               flow) includes a gas cooling tower, an inlet test duct with sampling
               port, a particle charging corona section (corona no. 1), a charged
               water spray tower (tower no. 1), a particle charging corona section
               (corona no. 2), a charged water spray tower (tower no. 2), a posi-
               tively charged corona section to collect the positively charged water
               droplets, an outlet test duct with sampling port, and a fan.  The
               pilot plant is housed in a 40 ft. long trailer and can be easily trans-
               ported to emission sources.  A photo of the pilot plant (Mark 2P model)
               located at a steel plant is shown in Figure 3.
                                           305

-------
                                                         •s
                                                         CM


                                                         J4
                                                          U
                                                          4J

                                                          5
                                                          
-------
                                                                          PH
                                                                          0)
                                                                          0)
                                                                         4-1
                                                                         c
                                                                         a)
                                                                         4-J
                                                                         o
                                                                         •H
                                                                         (X
                                                                        •§
                                                                         u
                                                                        •H
                                                                        tn
                                                                        o
                                                                        u
                                                                        01
                                                                        w
                                                                        (JO

                                                                       El
307

-------
      B.  Test Methods

          The particle size distribution and mass concentration were simul-
          taneously measured at the inlet and outlet test duct using UW Mark 3
          and UW Mark 5 Source Test Cascade Impactors.   During some tests  the
          water charge/mass and aerosol charge/mass  were measured.   The test
          parameter and measurement techniques are presented  in Table 1.
             TABLE 1.   Source Test Parameters and Measurement Techniques
                   Parameter
          Equipment
           1.   Air

               a.   velocity and volume


               b.   temperature

               c.   moisture



               d.   atmospheric pressure

               e.   static pressure

           2.   Water Spray Towers

               a.   water flow

               b.   water charge to
                   mass ratio

           3.   Aerosol

               a.   mass concentration

               b.   size distribution

               c.   aerosol charge to
                   mass ratio
S-type pitot tube with
draft gauge

thermometer

wet and dry bulb thermometer
and checked by volume of con-
densate

barometer

Magnehelic gauge
 rotometers

 digital multimeter




 UW Mark 3 or 5 Cascade Impactor

 UW Mark 3 or 5 Cascade Impactor

 digital multimeter
III.   Results

      A.   Dioctyl Phthalate

          In order to check out the newly constructed pilot plant,  eight tests
          were conducted using dioctyl phthalate (DOP)  aerosol.   The test re-
          sults are presented in Table 2.
                                      508

-------
                    TABLE 2.  DOP Test Results
Test
No.
1
2
3
4
5
6
7
8
Gas Flow at
Outlet Duct
(acfm)
744
533
335
515
480
959
969
970
Water to Outlet
Gas Flow Ratio
(gal. /I, 000 acf)
5.8
8.4
13.4
8.9
12.7
6.4
6.2
6.2
Voltage (KV)
Corona
#1
0
0
0
70
70
70
0
0
#2
70
70
70
70
70
70
0
0
Spray
#1
0
0
0
25
15
25
0
15
#2
25
25
25
25
15
25
0
15
Collection
Efficiency
(%)
92.1
94.7
99.7
98.1
95.5
89.0
25.0
49.8
Outlet
Cone .
(gr/scf)
0.0007
0.0019
0.0001
0.0024
0.0044
0.0040
0.0227
0.0201
B.  Coal-Fired Boiler
    Three tests were conducted on the emissions from the University of
    Washington pulverized coal-fired boiler.  The boiler was operating
    at about 67% of its rated capacity during the three tests.  Soot-
    blowing occurred during the second test.  The spray liquor used in
    the scrubber towers was fresh water (no recycling of the liquor).
    The test results are presented in Table 3.
          TABLE 3.  Results of Tests at Coal-Fired Boiler
Test
No.
1
2
3
Gas Flow at
Outlet Duct
(acfm)
997
974
1,060
Water to Outlet
Gas Flow Ratio
(gal. /I, 000 acf)
5.8
5.7
2.4
Voltage (KV)
Corona
#1
65
60
60
#2
65
60
60
Spray
#1
20
15
20
#2
20
15
20
Collection
Efficiency
(%)
99.5
98.1
96.1
Outlet
Cone.
(gr/scf)
0.002
0.007
0.013
C.  Electric Arc Steel Furnace

    The pilot plant trailer was modified by:   (1)  increasing the liquor flow
    rate capabilities, (2) installing a liquor recycle system,  (3)  improving
    the purge air supply, (4) installing a spark-rate controller to provide
    a higher operating voltage to the corona  sections, and (5)  revising the
    water spray banks.  Then the pilot plant  was connected to a duct ex-
    hausting from two electric arc steel furnaces.   A photo of  the  pilot
    plant trailer (the building in the background  is a baghouse)  at the elec-
    tric arc steel furnace location is presented in Figure 3.   The  test re-
    sults for this emission source are presented in Table 4.
                                 309

-------
TABLE 4.   Results of Tests  at  Electric Arc Steel Furnace
Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Gas Flow at
Outlet Duct
(acfm)
1,783
1,484
1,553
1,032
1,474
1,428
1,391
1,281
1,470
1,214
1,331
1,281
1,210
1,225
1,259
1,255
1,189
1,174
1,163
1,175
1,148
1,221
1,247
1,293
1,184
905
953
1,309
1,338
1,296
1,302
1,298
1,122
Water to Outlet
Gas Flow Ratio
(gal./lOOO acf)
10.4
12.1
11.9
17.0
12.6
17.5
18.0
19.5
17.0
20.6
18.8
19.5
19.8
19.6
19.1
0
21.0
21.3
21.5
21.3
21.8
17.2
16.8
16.2
17.7
23.2
22.0
16.0
15.7
12.3
11.5
11.6
13.4
Voltage (KV)
Corona
#1
70
35
70
70
70
70
70
50
50
50
70
70
0
0
0
70
70
68
68
70
0
70
70
65
70
70
70
70
70
70
70
70
70
#2
70
35
70
70
70
70
70
50
50
50
70
70
0
0
0
70
70
68
68
70
0
70
70
65
70
70
70
70
70
70
70
70
70
Spray
#1 #2
15
15
20
0
15
10
20
0
20
10
20
0
20
0
10
0
20
10
0
10
10
0
0
0
0
0
0
0
0
10
2
2
2
15
15
20
0
15
10
20
0
20
10
20
0
20
0
10
0
20
10
0
10
10
10
10
0
0
10
10
10
10
10
2
2
2
Collection
Efficiency
(%)
94.2
94.0
98.1
95.3
92.6
97.4
87.9
89.0
80.2
81.0
91.0
85.6
83.7
80.4
58.8
87.9
93
97
98
98
89
98.6
96.4
93.6
98.1
99.0
98.9
86.5
83.7
96.6
98.6
98.8
99.6
Outlet
Cone.
(gr/scf)
0.0057
0.0025
0.0024
0.0016
0.0395
0.0025
0.0075
0.0978
0.0750
0.0811
0.0430
0.0178
0.1042
0.16797
0.33031
0.07380
0.0441
0.0285
0.0269
0.0313
0.1151
0.0194
0.0258
0.0315
0.0074
0.00344
0.00524
0.15362
0.10095
0.00678
0.0100
0.00992
0.00458
                           310

-------
         IV.   Discussion of Results

              A.   Dioctyl Phthalate Tests

                  The particle collection efficiency is related to the gas residence
                  time in the electrostatic scrubber.  This is illustrated in Fig. 4,
                  where the particle collection efficiency graph for test no. 4 and
                  test no. 2 are compared (test no. 4 had a 37% longer gas residence
                  time than test no. 2).  Test no. 4 had a collection efficiency for
                  0.5 micron diameter particles of 98.5% compared to 91.5% for test no.  2.
Porticle
Collection
Efficiency
             95
             90
             85
             80
             75
Notes:    I. Gas  residence  time  in test 4
           approximately 37% longer than
           test 2.
        2. All  other pilot plant  operating   —
           parameters  essentially unchanged
                                                                                I  I I  I I h III!
               0.1          .2     .3   .4  .5 .6   3  I.          2.     3.   4.   5.  &    8.   10.

                                        Particle  Diameter  (microns)

              Fig.  4  Influence of Gas Residence Time on Particle  Collection Efficiency
              The particle collection efficiency at any given particle diameter  increases
              when either the particles or the spray liquor droplets change from uncharged
              to electrostatically charged conditions.  The particle collection  effici-
              ency versus particle diameter curves in Fig. 5 show that the collection
              efficiency for 0.5 micron diameter particles increases from 10%  in test  7
              (both particles and droplets uncharged) to 23.5% in test 8 (particles un-
              charged and droplets charged) to 91.5% in test 6 (particles and  droplets
              charged to opposite polarity).
                                             311

-------
Particle
Collection
Efficiency
                                        OOP Test  No. 6
                                        (particles  and droplets charged
                     OOP Test  No. 8
                     (particles  uncharged
                     droplets  charged
                                                         OOP  Test No. 7
                                                         (particles  and  droplets
                                                         uncharged)
                           .2
4  5  .6   8   I.          2.
  Particle  Diameter  (microns)
                                                                         3.    4.   5.  6.   8.  10.
           Fig. 5   Influence  of Aerosol  and  Droplet  Charging  on Particle Collection
                    Efficiency
                   Coal-Fired  Boiler

                   The  spray liquor charging  voltage affects the particle collection ef-
                   ficiency as is  shown  in Fig.  6.   Increasing the liquor charging volt-
                   age  increases the  particle collection efficiency at any given particle
                   diameter.   For  example,  test  2 (15 kV droplet charging voltage) had a
                   96.5%  collection efficiency for  0.5 micron diameter particles, whereas,
                   test 1 (20  kV droplet charging voltage)  had a corresponding 99.1% col-
                   lection efficiency.

                   The  influence of the  water to gas flow rate ratio is illustrated in
                   Fig. 7 where test  3  (2.36  gal./I,000 acf) had a 76.0% collection ef-
                   ficiency for 0.6 micron diameter particles while the corresponding ef-
                   ficiency for test  1  (5.82  gal./I,000 acf) was 97.8%.  Thus an increase
                   in the water to gas flow rate ratio provides an increase in the parti-
                   cle  collection  efficiency  at  any given particle size.

                                              312

-------
            100
            95
 Particle     90
 Collection
 Efficiency
    (%)
            85
             80
—  Coat-Fired  Boiler
-   Test NO. I
                                                      Coal-Fired  Boiler  Test No. 2
                                 Notes:   I. Water droplet charging voltage
                                           = 20  kV on test  no. I  and  15 kV
                                           on  test  no. 2.
                                        2. Alt other pilot  plant  operating
                                           parameters  essentially unchanged
               0.1
           .2      .3    .4   .5 .6   .8  I.           2.

           '             Particle  Diameter (microns)
                                                                          3.    4.   5.  6    8.   10.
          Fig. 6   Influence of Water Droplet Charging Voltage on Particle  Collection
                   Efficiency
            100
            95
Porticle     90
Collection
Efficiency
   (%)
            85
            80
                TTTTTnYrTTTTTI1   |  I |  I |  I I I  I IIII  i I  I'TITTl II  I III

                                                 •»——__
                               Coal-Fired Boiler  Test No. I
         Coal-Fired  Boiler  Test No. 3
                                       Notes: I.  Water  to  gas  ratio (gaL/
                                                  1,000 ocf ) = 5.82 in  test
                                                  nal  and  2.36  in  test
                                                  no. 3
                                               2.  All other  pilot plant
                                                  operating  parameters
                                                  essentially  unchanged
            75 II I  I I I I 11 II  I I  1 I I   I  I  I  I  I.I I 11 II ll I  I I  I I I I 11 il  I I  I I
               O.I
           .2      .3    4   .5  .6   .8   I.          2.      3.    4.   5.  6.    8.  10.

                        Particle  Diameter  (microns)
            Fig.  7  Influence of Water to Gas Flow Rate Ratio  on Particle Collection
                    Efficiency
                                             313

-------
The overall particle collection efficiencies listed in Table 3 do not
include the particles collected in the water spray cooling tower
located upstream of the inlet sampling port.  The inlet particle con-
centrations ranged from 0.3 to 0.4 grains/scf as sampled downstream
of the spray cooling tower.  Therefore, it is expected that consid-
erably higher overall particle collection efficiencies would be cal-
culated if the inlet sampling was conducted at a location upstream of
the cooling tower where the particle concentrations are in the 0.5 to
5 grains/scf range.  However, the particles removed in the cooling
tower are in the larger size range and thus particle collection eff-
iciency versus particle size curves presented in Fig. 6 and 7 should
not be significantly affected.

The test results for the electric arc steel furnace (Table 4) were
obtained during tests conducted from January to May 1977.  These are the
first tests with the new liquor recycle system.  Some problems occurred
during the earlier tests.  During tests 1 through 16, particle re-
entrainment was occurring from the duct downstream of the mist eliminator.
This caused the overall particle collection efficiency to be less than
expected for these tests.  The particle re-entrainment problem was
detected by a test performed with clean  (atmospheric) air which showed
a higher outlet particulate concentration than at the system inlet (clean
water was used as the scrubbing liquor).   Washing  down  the  duct down-
stream of the mist eliminator corrected this situation.

During tests 22 to 29, the liquor spray to tower 1 was shut off because
it appeared that the spray was flooding the no. 2 corona section.  After
test 29, the pilot plant was shut down and the spray towers and corona
sections were washed down thoroughly.  Of the six nozzles in tower 1,
the dowstream nozzle fittings were plugged, and the other three nozzles
were replaced with nozzles providing a fine mist (manufacturer data
specifies 200 to 300 micron diameter droplets).  A screen-type mist
eliminator was installed at the outlet of tower 1 (inlet to corona 2).
Also all the spray nozzles in tower 2 were replaced with the finer droplet
nozzles.  Tests 30 through 33 were run with both corona sections and
both spray towers in operation at a total liquor flow of about 16 gal/
minute and a liquor pressure of about 56 psig.

As shown in Fig. 8, the particle collection efficiency versus particle
size curves obtained at similar operating conditions are in fairly good
agreement.  Tests 26 and 27 were conducted with a lower gas volumetric
flowrate and correspondingly the corona section SCA and liquor to gas
collection efficiencies for these two tests (nos.26 and 27) were in excess
of 99%.

Fig. 9 presents the collection efficiency versus particle size for tests
31 through 33 and shows that the submicron particle collection efficiency
for these tests exceeds about 94%.  Also these tests illustrate the ability
to achieve good particle collection at lower liquor spray voltages and at
lower liquor to gas flowrate ratios.
                           314

-------
    99.9
O
z
UJ
O
UJ
z
O
O
UJ
8
UJ
_J
O
H
cr
2
99.0
90.0
           Test Corona V.
           No.    (kV)
                                   T
                                                         I i
            22
            23
            26
            27
            28
            29
                    Spray V.    Overall      SCA       L/G
                      (kV)   Coll.Eff.(°Q (ftz/cfm)  (gal/IOOOcf)
                                         TEST 28
                   ELECTRIC  ARC STEEL  FURNACE
                                                         I i i i
0.001
0.01
                                                                 ir
                                                                 »-
                                                                 UJ
                                                                 z
                                                                 UJ
                                                                 Q_
O.I
      0        .2       .4   .6  .8  I        2       4    6   8 10

         PARTICLE AERODYNAMIC DIAMETER, d5Q(microns)


        Fig. 8  Influence of SCA and L/G on Particle Collection Efficiencies

                                    315

-------
    99.9
o
z
UJ
o
u_
u.
LJ

O
I-
o
UJ
UJ
_l
o
tr
£
99.0
90.0
       Test Corona V. Spray V.   Overall        SCA      L/G
       No.    (kV)  '   (kV)   Coll.Eff.U)  (ft2/cfm)  (gal/IOOOcf)
            31
            32
            33
              70
              70
              70
98.6
98.8
99.6
.059
.060
.082
11.52
11.60
13.40
                            i  i  i  i
                                     I
                                                         i i  i
                                                              0.001
                               0.01
                                                                    IT
                                                                    UJ
                                                                    z
                                                                    UJ
                                                                    Q.
                               O.I
                .2        .4   .6  .8  I        2        4    6   8 10
         PARTICLE  AERODYNAMIC  DIAMETER, d50(microns)

         Fig. 9  Collection Efficiencies at an Electric Arc Steel Furnace
                                     316

-------
V.  Conclusions

     The results of tests of the UW Electrostatic Scrubber field portable pilot
plants illustrate the capability of high efficiency fine particle collection
efficiency at low gas pressure drop (about 0.5 inches of water)  and over a range
of liquid to gas flowrate ratios (2.4 to 21.8 gal/1000 acf).   Measured overall
particle collection efficiencies range from about 25 to 99.8% depending upon the
electrostatic scrubber operating conditions and the inlet particle size distribution.
Acknowledgements

     This research was supported in part by EPA (IERL)  Research Grants
No. R 803278 and R 8043930 and research grants from Reynolds Metals Com-
pany,  Alaska Lumber and Pulp Company, Ketchikan Pulp Company, Georgia
Pacific Corp., Scott Paper Company, ITT Rayonier Inc.,  and Publishers
Paper  Company.  The assistance of the Bethlehem Steel Corp., Seattle, is
gratefully acknowledged.
                                   317

-------
References
1.  Kraemer, H.F. and H.F. Johnstone (1955)  "Collection of aerosol particles
         in the presence of electric fields"  Ind.  Engr. Chem.  47 2426.

2.  Penney, G.W. (1944)  "Electrified liquid  spray dust precipitator"  U.S.
         Patent No. 2,357,354.

3.  Pilat, M.J., S.A. Jaasund, and L.E. Sparks (1974)  "Collection of aerosol
         particles by electrostatic droplet  spray scrubbers"  Envir. Sci.  &
         Tech.  8.  340-348.

4.  Pilat, M.J. (1975) "Collection of aerosol particles by electrostatic
         droplet spray scrubber"  APCA Journal  Z5  176-178.

5.  Pilat, M.J. and D.F. Meyer (1976) "University of Washington electrostatic
         spray scrubber evaluation"  Final Report on Grant No.  R-803278, EPA
         Report No. EPA-600/2-76-100 (NTIS No. PB 252653/AS).
                       Metric Conversion Factors

Readers more familiar with the metric system are asked to use the conversion
factors tabulated below:
  Non-Metric                Multiplied by:               Yields metric
      ft                       30.48
                                                               cm
      ft3                      28.32                          liter
                                                                 2
  grains/scf                    2.288                    1 gram/m

     gal.                       3.79                          liter

  gal./lOOO acf                 0.1337                      liter/m3

    ft /min                     1.699                        m3/hr

    ft/sec                      0.3048                       ra/sec
                                   318

-------
     FINE PARTICLE SCRUBBING WITH LONG STAR STEEL HYDRO-SONIC CLEANERS
                              "THE COALESCER"
                              Thomas  K.  Ewan
                               Jay S.  Master
                          Lone Star Steel  Company
                           Lone Star,  Texas  75668
                                ABSTRACT

     Lone Star Steel has developed a family of Hydro-Sonic air cleaners for
fine particulate control.  These cleaners can also control S02 and t^S with
chemical addition and particulate in a single system.

     Cleaners are available in a number of different models, sizes and  mate-
rials to meet varied needs of differing processes.

     These cleaners are powered by either:  (1) an I.D. or F.D. fan drive,
(2) a compressible fluid ejector drive, or (3) combinations of both.  The
energy range is from <1 BTU per pound of off-gas to 100 BTU per pound of
off-gas depending on the pumping and cleaning requirements of the process.
The precise amount of energy can be inputted to within 1 BTU per pound of
off-gas for a specific cleaning level.  Selective particulate size removal
can be carried out.  The cleaning level is adjustable by varying the energy
input.  The capital equipment does not have to be changed to effect this.

     Aerodynamic technology is employed for high performance cleaning com-
bined with minimum energy utilization:

          1.  "Free Jet" and shock wave mixing.

          2.  "Free Jet" formation of droplets shaped into "paddle"
              and flat-like surfaces.

          3.  Shaped directional cleaning energy release for fine
              particulate control also controls opacity.  The
              ejector driven models release the total energy of
              the system through a narrow cone concentrating clean-
              ing forces.

          4.  The separation function is performed either by cy-
              clones or by energy recovery modified diffusers.

     These cleaners have been under development since 1969 by a R&D team
composed of engineers and physicists.  We have five years of continuous
data taking and development.  While the amount of energy inputted is im-
portant, we have learned how to achieve high level cleaning by the manage-
ment and application of that energy in order to achieve high performance
and economy at the same time.

                                   329

-------
       FINE PARTICLE SCRUBBING WITH LONE STAR STEEL HYDRO-SONIC CLEANERS

                                "THE COALESCER"
Introduction

     Lone Star Steel Company has developed a patented family of Hydro-Sonic
air cleaners for control of particulate, SO^, H^S, and precursor gases.
These include the compressible fluid driven models (steam or air) or the
fan driven models or combinations thereof employing either cyclone or dif-
fusing coalescers for the separation function.  These cleaners cover a
broad range of cleaning capabilities from meeting regulatory standards to
virtual elimination of pollutants.  The selection of the cleaner to be
employed is made based on the application, the cleaning requirements, and
the source of energy available.

     While the cleaning performance for the Hydro and Coalescer are the
same for particulate, the Coalescer cleaner (a device embodying aerody-
namic techniques for cleaning and energy recovery) is discussed in this
paper.  The Coalescer combines high performance with energy recovery in
a single device for the removal of particulate and water soluble gases.

     High performance with low energy requirement is achieved by the use of
a "Free Jet" mixing stream and water drops of the proper shape, size, number
and velocity combined with the use of a specially designed pressure recovery
separator (diffuser).  The special diffuser is used to provide separation of
the cleaned gases from the particulate-containing water and to recover pres-
sure by transforming a portion of the velocity head of the gas stream into a
pressure head.  The Coalescer Separator, in common with cyclones, demisters
or other separation devices performs a separation function, but unlike other
separators, the Coalescer recovers a portion of the pressure thereby mini-
mizing the net energy requirement.

     The Coalescer system may be powered by a fan or a compressible fluid
ejector or combinations thereof.  (Figure 3 - Fan Coalescer.  Figure 4 -
Ejector Coalescer.)  The energy requirement for the Coalescer system is low
when compared to conventional wet scrubbing for comparable cleaning perfor-
mance.

Discussion

     The Coalescer - How It Works.   The operation of the Fan Coalescer and
the Ejector Driven Coalescer will be discussed separately.

     The Fan Coalescer is a subsonic device in which all of the process off-
gas is pushed by a fan through a subsonic nozzle, Figure 3.  The operating
nozzle pressure is from 2" to 50" of water depending on the process to be
cleaned.
                                      320

-------
     The Ejector Coalescer is a supersonic compressible fluid drive with no
moving parts in the gas stream.  The compressible fluid is preferably air or
steam.  It operates with 15 PSIG to 1000 PSIG or higher driving pressure de-
pending on the available steam or compressed air source.

     In the Coalescer cleaners particulate is wetted and the water drops
containing particulate are grown to an easy-to-remove size.  The mechanisms
employed are simple: first, the wetting of the particulate; second, the
complete mixing of the particulate and the scrubbing medium; third, the
process of growth; and fourth, the process of separation.  The processes of
wetting and mixing in a "Free Jet", combined with shaped droplets, droplet
growth, and separation with energy recovery, distinguish these cleaners from
conventional systems.

     The wetting of the particulate is achieved initially by the generation
of very large numbers of high speed water droplets of the proper shape, num-
ber, size and speed to contact the particulate and/or noxious gases.

     Figure 5 shows the generation of water droplets of the proper shape,
size, speed and number with the subsonic drive while Figure 6 shows the
supersonic drive.  First, a water droplet is mechanically produced which
droplet is in turn injected onto the exhaust of the off-gas stream or onto
the compressible fluid ejector stream.  When the presized water droplet
strikes the high speed stream, operating in a "Free Jet", it is shattered
into many smaller droplets and accelerated to high speed.  It is desirable
to create high speed droplets of a size 10 to 30 times the size of the par-
ticulate to be captured since the dust particle may be able to go around or
dodge an over-large water drop.  The key to the capture of fine particulate
and gases is, therefore, the creation of water droplets of the proper shape,
number, velocity and size for the particular gas stream.  In the "Free Jet"
arrangement, it is believed that the water droplets are formed into paddle-
like or flat surfaces before contacting the particulate.  This increases the
surface contact area and assists in direct contact between the particulate
and the water.  Subsequently, it is believed that the surface tension exerted
on the misshapen water droplet causes the droplet to close around and encap-
sulate the dust particle.

     The mixing process immediately follows and overlaps the wetting process.
It insures complete wetting of any particles not already wetted by collision.
Mixing is accelerated and enhanced in a "Free Jet" as compared to mixing in
a confined area or region.

     The growth process follows and partly overlaps the mixing process so that
a single water drop may contain hundreds of micronic and submicronic dust par-
ticles.  As a result of the growth of droplets containing particulate into
increasingly larger size, the initial size and shape of the particulate in the
gas stream has only a small effect on its removal.

     Figure 7 illustrates droplet growth.  A Bertin instrument (France) is
used to capture a sample of the mixture in the gas stream.  On the left is
shown what appears to be a large particle "captured" in a water droplet;

-------
however, this same droplet observed under a microscope after evaporation
exhibits the condition shown on the right side of the picture, where doz-
ens of tiny particles appear.  This is confirmation of the coalescence of
many small particulate-containing droplets into large droplets.  The
ability to grow particulate-containing droplets into 1000 micron or larger
size is important in high performance low energy particulate abatement since
separation and removal of such large water drops is easy.

     Separation of Cleaned Gas and Particulate-Containing Water Droplets.
The separation of the cleaned gas from the water droplets containing the
particulate is achieved in a modified diffuser by the use of detached flow,
directional forces, velocity control, coalescence and gravity forces at the
same time the diffuser recovers pressure.  Figures 3 and 4 illustrate sche-
matically the operation of the modified diffuser.  There are three control-
ling interactions—pressure, velocity and direction.  Normally in fluid
flows and aerodynamics, a diffuser is used for pressure recovery, i.e., to
reduce the energy required in a system.  The modified diffuser of the Coales-
cer cleaner provides partial pressure recovery, but its primary function is
to separate the particulate-containing water drops from the gas.  The modi-
fied diffuser thus replaces cyclones, stilling chambers, impactors and/or
mist eliminators for the separation function while at the same time providing
the added benefit of pressure recovery.

     At the entrance to the modified diffuser, the droplet normally moves at
a speed in excess of 100 feet per second.  At this point, the flow of the
gas and water mixture is partially separated (detached) from the diffuser
wall so as to direct the particulate laden water drops toward the bottom
and away from the cleaned gas.  The top of the modified diffuser (as shown
in Figures 3 and 4) includes a region of lower pressure which causes re-
circulation and reentrainment of small droplets and mist.  The small drop-
lets are thus returned to the coalescing zone of the diffuser where rapid
droplet growth occurs.  This process augments the earlier mixing and growth
process.  Water and particulate separated from the mixture are removed
through the drain by gravity forces while the cleaned gas exits in the up-
ward (opposite) direction.

     In addition to pressure control and droplet directional control discus-
sed above, a third phenomenon occurring in the modified diffuser is droplet
growth as regulated by velocity control.  At the entrance to the modified
diffuser, the speed of the droplet is usually greater than 100 feet per
second.  At the exit end of the modified diffuser, the velocity has been
reduced by a factor of 5 to 10 because of the increased cross sectional
area of the diffuser.  Massive coalescing (droplet growth) results from
this rapid velocity change.  At this point, the particles are encapsulated
in water drops grown to a size easy-to-remove by gravity.

     The Coalescer cleaners do not seek to remove submicronic and micronic
particulate as such, but instead wet the particulate and then grow the
droplets containing the particulate to easy-to-remove size.

                                       322

-------
     The Coalescer may be driven either by a fan or a compressible fluid
ejector nozzle.

     The Subsonic Fan Drive is used to drive the entire off-gas from the
process through a subsonic nozzle.  It is used in applications where a fan
will: provide trouble free service in the off-gas stream; and where the
driving energy required is less than 50" HoO W.C. at the subsonic nozzle.
This drive pumps the gas stream and provides the cleaning energy.  It meets
the needs of most applications where cleaning requirements are in the .OX
range.  Energy to pump and clean one pound of off-gas, depending on the
process and the required draft, is in the <1 to 15 BTU range.

     The Supersonic Ejector Drive is used in applications where: it is
desirable and/or necessary to have no moving parts in the gas stream; the
energy requirement would be greater than 50" W.C. with a fan drive using
a subsonic nozzle; there is a hostile environment or very difficult to
remove gases or particulate; the standards are highly restrictive; there
exists significant quantities of particulate in the submicronic range; and
where cleaning is desired in the .OOX to .OOOX grains per SCF outlet.  The
energy to pump and clean one pound of off-gas, depending on the process and
the required draft, is in the 15 to 150 BTU range.

Performance of the Ejector Coalescer

     In the compressible fluid Ejector Coalescer, the pumping of the gas
stream and the energy to clean the stream are both furnished by a supersonic
ejector nozzle.  The supersonic nozzle is normally driven by air or steam at
a pressure of 15 PSIG or greater.  15 PSIG is necessary to achieve supersonic
flow.  Some steam driven units are operating in the 500 PSIG range; however,
this is not an upper limit.  When supersonic flow exists, a series of shock
waves is established which aids in the mixing process and insures complete
wetting of any particles not already wetted.  The shock waves form violent
scrubbing zones where intimate contact is forced to occur between the scrub-
bing medium and the particulate and/or noxious gas and provide the added
contact and mixing that make the cleaner superior for fine particulate and
SC>2 removal.  When using the air drive, the Ejector Coalescer can also providc-
the oxidizing conditions desired for certain processes, such as S0~ control.
Ejector models are in full scale operation on ore kilns and open hearth fur-
naces and successful pilot demonstrations have been conducted on a number of
other processes.

Performance of the Fan Coalescer

     Unlike the ejector models, the Fan Coalescer does have moving parts in
the gas stream.  This is acceptable in many applications.  The performance
of the Fan Coalescer has been demonstrated on various processes including
coke ovens, sinter plants, ore kilns, galvanizing mills and cupolas in the
steel industry.  In the paper industry, the Fan Coalescer performance has
been demonstrated on recovery boilers, a bark boiler and lime kilns.   As
shown on the following pages, low energy requirements are combined with ex-
cellent performance.  This is possible because of the "Free Jet" mixing,
droplet shape and other features described earlier.

                                     323

-------
     Performance on specific processes is discussed below.  Inasmuch as an
open hearth furnace (especially during high heat and oxygen lance) produces
extremely large quantities of very fine, difficult to capture particles, it
is used as typical of the collection efficiency of the LSS ejector driven
Hydro-Sonic cleaners.  Other processes discussed in this paper are less dif-
ficult to clean and are presented with inlet grain size information vs clean-
ing performance in grains SCF outlet.

Open Hearth

     The ejector drive Hydro-Sonic cleaner on an open hearth furnace includ-
ing oxygen lance provides controllable cleaning from .OX to .OOOX grains
(partial) per SCF outlet range.  Table I shows the scrubber inlet grain
loading and the size distribution.  Table II shows the scrubber outlet grain
loading and size distribution.

     Figure 8 shows collection efficiency vs particle size (EPA Report 650/
2-74-028).}

Electric Furnace

     The particle size for an electric furnace (as reported in NTIS PB-
203521)2 is shown in Table III.  LSS Hydro-Sonic cleaners consistently
provide cleaning well below the .0052 grains SCF outlet EPA regulation on
the LSS electric furnace.  This is achieved with less than 50 BTU per
pound of off-gas pumped and cleaned.  (See Figure 9.)

Galvanizing

     Table III exhibits the inlet particle size distribution for the pro-
cesses discussed below.

     Zinc dust is readily controlled by use of the low energy Fan Coalescer
as shown in Figure 10.  The effectiveness of this cleaning on this particu-
late is such as to provide a clear stack and outlet grain loading many times
within standards.  The capture of this particulate makes possible the re-
covery of valuable materials from an off-gas stream.  It is being used at
Lone Star Steel for the large scale recovery of valuable metals from its
open hearth furnaces, in addition to the galvanizing process.  The quantity
of valuable metals recovered from the galvanizing line is such that a net
profit is realized over the cost of cleaning the air.  In general, it has
been found that as the particulate becomes smaller the concentration and
recovery value of the metallic component increases.

Sodium Sulfate Recovery Boiler

     A sodium sulfate recovery boiler is readily cleaned using the Fan
Coalescer, Figure 11.  It provides cleaning well within required standards
plus the added feature of capturing and returning to the process valuable
material contained in the off-gas.  As in the case of the galvanizing mill,
a Coalescer on a recovery boiler can, by the capture of fine particulate,

                                      324

-------
reduce the cost of cleaning the off-gas; and, in some applications, provide
a net profit over the cost of cleaning.  Figure 11 shows the cleaning perfor-
mance of the Fan and Ejector Coalescer compared with the performance of a
Steam Hydro cleaner on the same pulp mill recovery boiler.

Sinter Plants

     Sinter Plants are one of the more difficult processes to clean, but the
Fan Coalescer is capable of cleaning sinter off-gas to the degree shown in
Figure 10.  If large quantities of hydrocarbons are present or if more strin-
gent cleaning requirement is required, use of the air or steam Ejector Coales-
cer rather than the Fan Coalescer is recommended. However, the economics of
the Fan Coalescer make it the choice for sinter plants using coke-breeze to
meet existing and proposed regulations.

Cupola

     Cupolas are used to melt down many forms of "waste" and scrap.  Engine
blocks, mill scrap, transmissions, and other scrap materials, many of which
contain significant amounts of oils and hydrocarbons, are often used.  The
Fan Coalescer has demonstrated satisfactory cleaning performance on cupolas
charged with these materials, Figure 10.  In addition to meeting the grain
loading requirements with significantly less energy than the so-called
"conventional wet scrubbers", the Fan Coalescer will clean a cupola so as to
produce a "clear stack".

Coke Oven

     Figure 10 shows the performance of a Fan Coalescer in cleaning the off-
gases of a coke oven.

All Particulate Processes

     Figure 12 shows that the LSS Hydro-Sonic cleaners have the capability
to clean any particulate of any size and any inlet grain loading to meet or
exceed published air cleaning requirements with reasonable economics.  The
precise amount of energy required can be inputted.  This is adjustable and
can be provided on a need required basis to meet changing conditions.

Conclusions

     The Coalescer represents new technology in the air cleaning field.  It
provides high performance, low capital cost and low operating energy require-
ments.   A choice of driving mediums is available: either a fan or a compres-
sible fluid ejector may be used.  Normally, steam or air is employed as the
compressible fluid with the ejector model.  In the ejector models there are
no moving parts in the gas stream, and the ejector both pumps and cleans.  On
both the fan and ejector models, particulate, S02, and ^S may all be control-
led in a single unit.  The desired cleaning level may be modified by a change
in energy supplied to the ejector or fan.   In the fan models, an increase in
nozzle pressure will provide an increase in cleaning performance.  In the
ejector models, an increase in the amount of steam or air in relation to the
off-gas being pumped and cleaned will provide increased performance.

                                      32 5

-------
     Lone Star Steel tests with the ejector drive have demonstrated highly
effective removal of hydrophobic fumed silica having a near uniform particle
diameter of .007 microns.  This material rejects water.  Analysis of the
captured material shows the particulate not wetted, but encapsulated in a
film of water.  Encapsulating of the dust particle in water is one of the
important mechanisms for the capture of fine particulate in the Coalescer
cleaner.

     These cleaners operate at relatively high gas thru speeds so that
residence time in the contact and mixing zone is measured in milliseconds.
This allows the building of units of a small size and also provides for
savings in the chemicals used to clean off-gas streams containing I^S, CC>2
and the like.   The short residence time allows, for example the removal of
H2S by caustic, but does not permit sufficient time for the C02 which may
be present to react with the caustic.  The Coalescer provides a large num-
ber of small high speed water droplets having very high total surface area
which enables the cleaners to make instant contact between the droplets,
the chemical additive, and the gas to be scrubbed.  On all of the processes
tested to date, the Coalescer provided excellent performance with relatively
low energy.  In the Coalescer, a pressure recovery diffuser separator re-
places cyclones, demisters, impactors, and stilling chambers.  The Coalescer
combines air cleaning with pressure recovery—a new and economically advan-
tageous approach in air cleaning.
                                      326

-------
















to
CO

rl
01
c
•H
0)
N
•H
K

_Q

60
•H
•a
CO
o


4J
01
rH
B
•H
rl
01
,0
,0
3
rl
CJ
cn



<
M

01
rH
/•>
CO
H

































•4H
CJ
w
CO
c
•H
CO
M

rH
CO

tl
CJ
4-»
fS

CJ
N
-H
C/J
C
•H

6C
C
•H

CO
O
rJ































m ri
B 2

6-1 <














co"
CN

V
3
.
V
(
co en
CM  CM ^- ^J ON CJN rH ON LO r*- CM CM ^^ 00 "-3" rH |
in rH in r^- m NO O *vT ON r^- co ^ ^f \o CM CM r^ I-H \o CM CM A



co NOinco oo P^CM oo-3"ON

O rHOO CM CO^ i-HrHCM rH
. . . . . . . . . A





CO rH
•H II
° -3




\o •<]• vo i^". oo ^o r^«
SSooSSKSrS^IS^SSSrHcSocOrHSPiS
OrHrHOOCMOOOr-'CMCMrHCMCMrHCMrHCMrHrHO



Sp^SoS§rHi30rHcMcoco5inInin3cooooo'^





O Td 0)
OO CO rH P^ CM NO »^- in O ^ CM 00 P^» VO in CO in 00 ON P^ VO CM
rHin>
CO

0
4J
T)
01
01
13

0

01
CO
0)
01 01
i-H 60
Q. rl
e a
5 J3
en cj
CO ,£J
































































CO
4J
01
0
4-1
0

0
4J

01
60
rl
cti
u
MH
o

•o
B
J 1








































































01
0
c
CO
rH
rH
CO
4J
01
e

4J
O
S3


















































































a.
CO
IT
327

-------

















,_,
CD
>
CU
C
•H

CU
N
•H
CO

r>
"^
00

•H
•a
C8
o
r-\

4J
HI
rH
4-1
3
O

CU
43
43
3
rl
O
C/3


,
M
M

CU
j—l
n
CO
EH




























^
>~H

.
U
QJ

o
rH

t
O
O
Q



O*v

..
O
o
a





00


o
cu
Q





CO


,
O
cu






VD

,
O
cu
Q






vD

,
o
D





























*^_
LT) O
-Cf" rH
.. ..
O r-<


•v,^
LT) U"l
rH rH

rH rH



-^, -^^
O LO in o
co CM mo

O rH rH rH





-^
o o
en en
" •'
rH rH






•—^ — ^^
01 en m o
O 01 01 O

OH 01-3-
rH rH rH rH





*"•» r-- "*»^
m in in ex]
^ rH • <• en
.. .. o •• ••
rH rH Q rH rH






~_^
en en
O 01
•• ••
rH rH













lf^_
CO
••^ '
Q)

H

^j

CO

cu
4->
CO
P

O
o
o
o
m en rH


O
0
O
0
CO CO rH



o
o
0
o
CN CXI rH





O
0
O
O
01 01 rH






O
O
o
o
in rH
O1 CXI l— 1





o
o
o
o
in rH
01 CXI rH






o
o
o
o
01 CXI 1-4












cu
4-1
CO
Pi
cu
CMS
f"t "< ?\
U rJ U
•H tO r-l
4-1 CO tn
•H CD
CO M rl
(2 ^ 42
G 42
• co 3
CJ CU rl
CJ 4J CJ
 00
O t^-'









































































00
e
•rl
•a
co
o
rJ

'cfl
>
rl
CU
4-1
M

CU
N
•iH
CO













































































^-^
MH
CJ
CO

M
00

— -
(3
0

O
•H
S
m 0 as
rH rH O
0 O 0
0 O O
o o o
o o o

<• 
CO 00 O
a a
CU *iH
00 -O CU
tj CO rl
CO O (U
rH rH |S
C rH CO
•H CO CU
4-1 00
00 O CO
e 4-1 4-1
•H CO
4-1 CU
rH 43 rl
3 4J O
CO 4-1
CU CO O
rl 4-1 CO
a a,
•a cu S
CU CO iH
•H M >H
CO O.rH
MH CU ^
rl
CU 0) •
4J 3 CU
CO rH 00
cu at co
prj p> jj
CO rQ 10
328

-------















c
0
•H
3
•H
^
4J
CO
•H
T3
01
N
•H
Cfl
CD
CO
rH
3
O
•H
1 i
J-J
CO
O,
4j
0)
T-H
c
M
M
M
rH

CD
rH
X
CO
EH






















_
y
£ 2
O jyj
U
w Si
en £

°



6\






<" ^
•33
en




rH^
M
£*•»
QJ
/•^ ^
T O
^


o





01 ^
•H 3
en





2 S
~ rH
*H
Zj °
3 «
W >•>
| CD
•H >
21
«*>







CD _^^
en ^



M
C
•H
N v
•H g
rj S*
ed
£> ^
"rfl °
0 ^

en
en
^_]


0
o\






01
•33
en




o
j-i
 o^
• • • •
r^ r\] vo f\J
CN CM CN CN

in
•
en rH
o i i m
• in in r^-
cn • r^ •
V rH -V





in r^ en
en en CM







rH in
en l l
1 m rH




*sO
O 0\
r—j vO • •
vO r— f r-«. -
0
_f~;
CO
a;
N
•rH
CO

4-1
CO
0)
00
1-1
et)
rH
CU
C
cfl
4-1
1-J
0)
4J
cfl
0)
^4
GO
CO
«,-(

Jj
0>
a
CO
e
0)
M

CD
4J

o"
0
rH
c
CO
CO
Cfl
CD
rH

CO
•H

CO
CD
60
CO
C
CD
O
CD
a
CD
4J
O

B
3
CO
01
^J
CD

CD
§
329

-------
         SUPERSONIC
       EJECTOR  DRIVE
          SUBSONIC
         FAN   DRIVE
                                            COALESCER
           Figure  I.    Hydrosonic  cleaners.
         SUBSONIC
     SUPERSONIC
PREDOMINANTLY  SPHERICAL
CONCENTRATED  CONICAL
        Figure  2.   Energy  utilization  region.

                          330

-------
                           CLEAN GAS
               SUBSONIC  NOZZLE
OFF GAS
                        WATER INJECTOR
                      MODIFIED DIFFUSER
              Figure  3.    Fan  coalescer .
                            CLEAN GAS
               EJECTOR DRIVE

                         WATER  INJECTOR
                      MODIFIED  DIFFUSER
                                   • <-O-r/x'  ''   ^***f^
                                   ^^JL.
  OFF
          Figure  4.    Ejector  coalescer
                         331

-------
   Figure 5.  Shaped  droplets-subsonic  nozzle.
TSTEAM  OR   AIR
(] EJECTOR   DRIVE
   Figure 6.  Shaped  droplets-supersonic nozzle.
    IMMEDIATELY  AFTER      AFTER   EVAPORATION
          CAPTURE

   Figure 7.   Droplet   growth.
                          332

-------
g

H-
t;

UJ
Q.
        LONE  STAR  STEEL CO. OPEN  HEARTH
    0.01
     O.I
     1.0
            ACC.  STEAM  GAS
            POS. PRESSURE FLOW

          •  3    250    11000
          o  2    250    15000

          +  3    300    13000
                                 o
                                       +
           i  i  i i i i , ii °  i  i  i i . 11 ii
      0.01
                                              99.999
                                  99.99
                                 99.9
                               10.0
                                 99
                   O.I           1.0

                PARTICLE DIAMETER, //m

         DATA  RECORDED  BY SOUTHERN  RESEARCH

         INSTITUTE  FOR  E.P.A. — DEC., 1973


     Figure 8.     Supersonic  ejector  drive.
o


o

u.
UJ

z
o


UJ
                                        o
                                        o
     LONE  STAR  STEEL CO.  ELECTRIC  FURNACE
    CO

    o
    — Lu

    CO Q
    h- -E
    UJ O
    =)
    O
          .001
20  24  28  32  36  40 44  48
      BTU  LB. OFF GAS
     Figure  9.
     Coalescer - energy  vs. cleaning.

              333

-------
             LONE  STAR  STEEL  CO.  PROCESSES
           10      20     30      40     50      60     70
          NOZZLE   DRIVING   PRESSURE - inches  W.C.

Figure  10.    Fan  coalescer -  energy vs. outlet grain  loading.
                             334

-------
       FAN  NOZZLE  PRESSURE  (PN)~  INCHES  OF  H20
      5          10         20     30       50          100
o
CO
V)
c
'o
CO
o   -I
CO
CO
LJ


h-
UJ
ID
O
.05
   .03
   .02
   .01
^ 1 1 1 1 \
*. \
— . \
- '. \
'. \
• y
• ^
. \
\
• V
•
\
• \

%y
^r
\
•
\
».^
*,
*



-
•—
-
_







\
\
\

\
\
\

\
\
\
\ \
\
\ \
to \















AIR
1














HYDRO
1 III
-













a
^ \ \S' AIR COALESCER
\ .
\ >
\
•
"•A \
c
*'^
\
\
\

*
•^
*
\
STEAM HYDRO a __^\
STEAM COALESCER
(30 PSIG STEAM)
i I I I


(30 PSIG AIR)
.
V
i D
>\\D
\\ n
\ c
D\
N







»
*.




]


-
—
-
FAN COALESCER -r
', V
\ ^
X
/
^ \ /
Q \ /
\ *
\ \
^ \
\ '
\
,\



\
\
\
\
|Nxl 1 1
    .05          .1          .2     .3       .5          1.0
     EJECTOR  DRIVING  MEDIUM  (ex)  |b. / Ib. OF OFF GAS
         Figure II.
                         Paper recovery  boiler

                           335

-------
   120
O)
<
o


U-
LJ_
o


Lu
o
o
CL
100
LJ
_l
O
m
 80
    60
 40
    20
    0
           SUPERSONIC   EJECTOR   DRIVE
 MEDIAN
/ / /  /  /
        CLEAN   FROM  .OX TO  < .005  GRAINS/SCF.
        SUPER-SUB   EJECTOR/FAN  DRIVE

                      (EXPERIMENTAL)

           .CLEAN  FROM .OX TO <.005  GRAINS/SCF
                  SUBSONIC  FAN   DRIVE

                  CLEAN  FROM  .OX  TO .005

                                GRAINS /SCF
      0
                 2345

             PREDOMINATE  MICRON  SIZE
   Figure  12.   Energy  vs.  particle  size,
                        336

-------
                                  REFERENCES
1.    J.  D.  McCain and W.  B.  Smith,  "Lone Star Steel  Steam-Hydro  Air
     Cleaning System Evaluation",  Environmental Protection Technology
     Series,  EPA-650/2-74-028 (April 1974).

2.    L.  J.  Shannon,  P.  G.  Gorman and M.  Reichel,  "Particulate  Pollutant
     System Study Volume  II  - Fine  Particle  Emissions",  National Technical
     Information Services  Report,  PB-203 521 (August 1971).
                                    337

-------
                                 SYMPOSIUM ATTENDEES
J. ABBOTT
I.E.R.L. - MD-61
U.S. Environmental Protection Agency
Research Triangle Park, NC  27711

P.P. ADAMS
International Paper Company
P.O. Box 16807
Mobile, AL  36616

B. AGGARWAL
Ametek Process Systems
3427 Industrial Drive
Durham, NC  27704

F.J. ANDERSON
Carbon Products - Union Carbide
P.O. Box 6087
Cleveland, OH  44101

G. ANDES
AMAX Envir. Services Group
4704 Harlan Street
Denver, CO  80212

R. BACCAGLINI
ECAR Products Inc. (A-PAR)
436 W. Main Street
Wyckoff, NJ  07481

H. BARBARIKA
Air Pollution Technology, Inc.
4901 Morena Blvd., Suite 402
San Diego, CA  92117

E.T. BARROW
Ministry of Environment
880 Bay Street, 4th floor
Toronto, Ont, Canada M5S 1Z8

F. BASER
N.L. Industries
P.O. Box 1090
Highfstown, NJ  08520

R.A. BEANE
Air Pollution Control Commission
1558 Washington Street E.
Charleston, WV  25311

D.E. BECKNELL
Lear Siegler. Inc.
74 Inverness Drive,  E.
Englewood, CO  80110
D. BERGQUIST
Carbon Products - Union Carbide
P.O. Box 6087
Cleveland, OH  44101

C. BILLINGS
Environmental Engineering Science
740 Boylston Street
Chestnut Hill, MA  02167

R.C. BINNING
Monsanto Research Corporation
P.O. Box 8, Station B
Dayton, OH  45407

S. BIONDO
Federal Power Commission
825 N. Capitol St. NE
Washington, D.C.  20426

R.H. BOLL
Babcock § Wilcox Co.
P.O. Box 835
Alliance, OH  44601

J.W. BRADLEY
Griffin Pipe Products
2000 Spring Road
Oak Brook, IL  60521

R.M. BRADWAY
GCA/Technology Division
Burlington Road
Bedford, MA  01730

I.L. BURGENER
Pfizer, Inc.
2001 Lynch Avenue
E. St. Louis, IL  62201

W.R. CADY
Allied Chemical Corp.
P.O. Box 1139R
Morristown, NJ  07960

S. CALVERT
Air Pollution Technology, Inc.
4901 Morena Blvd., Suite 402
San Diego, CA  92117
I.E. CAMPBELL
Smelter Control Research Assn,
150 E. Broad St., Suite 601
Columbus, OH  43221
Inc.
                                         339

-------
P.R. CANDLER
Allied Chemical Corp.
P.O. Box 226
Geismar, LA  70734

K. CARLSSON
AB Svenska Flaktfabriken
Pack
S-35187 Vaxjo, Sweden

B. CARPENTER
Research Triangle Institute
Research Triangle Park, NC  27709
G.K.C. CHEN
Monsanto
800 N. Lindbergh Blvd.
St. Louis, MO  63166

G.H.S. CHENG
Environeering, Inc.
7401 N. Hamlin Avenue
Skokie, IL  60076

H.B. CHU
Department of Water § Power
111 N. Hope Street
Los Angeles, CA  90051

J. CLEMENTS
Standard Havens, Inc.
8800 E. 63rd
Kansas City, MO  64133

W. COLTHARP
Radian Corporation
8500 Shoal Creek Blvd.
Austin, TX  78766

F. CROWSON
Naval Surface Weapons Center
Code DG-31
Dahlgren, VA  22448

E.H. CUMPSTON
ECAR Products, Inc.
160 BenMont Avenue
Bennington, VT  05201

P.A. CZUCHRA
FMC Corporation
1800 FMC Drive West
Itasca, IL  60143
J.D'ANGELO
FMC Corporation
P.O. Box 4111
Pocatello, ID  83201

T. DEMBINSKY
C.A.P.C.C.
1 Penn Plaza
New York, NY  10001

D.C. DREHMEL
I.E.R.L. - MD-61
U.S. Environmental Protection Agency
Research Triangle Park, NC  27711

D.M. DUNKLE
Betz Laboratories, Inc.
Somerton Road
Trevose, PA  19047

G. DURFEE
Wyman-Gordon Company
244 Worcester St.
N. Grafton, MA  01536

L.A. DUVAL
Colerapa Industries, Inc.
4017 Hanway Blvd.
Ravenna, OH  44266

D.S. ENSOR
Meteorology Research, Inc.
464 W. Woodbury Road
Altadena, CA  91001

T.K. EWAN
Lone Star Steel Company
Highway 259 South
Lone Star, TX  75668

P.S. FARBER
Argonne National Laboratory
9700 S. Cass Avenue
Argonne, IL  60439

J.L. FASSO
St. Joe Minerals Corporation
Lead Smelting Division
Herculaneum, MO  63048

J.D. FERRELL
Steams-Roger Inc.
Box 5888
Denver, CO  80217
                                        340

-------
T.A. FITZPATRICK
E.I. DuPont 5 Co., Inc.
1 Golden Acres Drive, No. 8
Wilmington, DE  19809

E.W. GAGAN
APC Directorate
Government of Canada
Ottawa, Ontario, Canada K1A 1C8

J. GALESKI
Midwest Research Institute
425 Volker Blvd.
Kansas City, MO  64110

S.A. GANDHI
Air Pollution Technology, Inc.
4901 Morena Blvd., Suite 402
San Diego, CA  92117

W.J. GILBERT, JR.
Croll-Reynolds Co. Inc.
751 Central Avenue
Westfield, NJ  07091

J.R. GOSNEY, JR.
Owens-Corning Fiberglas
Fiberglas Tower FF/7
Toledo, OH  43659

R. GRANT
Department of Commerce
E. GRASSEL
Donaldson Co., Inc.
1400 W. 94th St, P.O. Box 1299
Minneapolis, MN  55440

G.P. GRUBER
E.I. DuPont § Co.,  Inc.
P.O. Drawer 219
New Johnsonville, TN  37134

G.F. HAINES
Bethlehem Steel
Research Department
Bethlehem, PA  18016

D.L. HARMON
I.E.R.L. - MD-61
U.S. Environmental  Protection Agency
Research Triangle Park, NC  27711
  H.N. HEAD
  Bechtel Corporation
  P.O. Box 3965
  San Francisco, CA  94119

  R. HERBEST
  Department of Commerce
  P. HUBBARD
  Ducon Company
  147 E. 2nd Street
  Mineola, NY  11501

  R.B. IDA
  U.S. Environmental Protection Agency
  Bldg. 53, Box 25227
  Denver Federal Center
  Denver, CO  80225

  R.J. JAWOROWSKI
  Air Correction - U.O.P.
  Tokeneke Road
  Darien, CT  06820

  H. N. JENKINS
  Jenkins S, Black, Inc.
  P.O. Box 80609
  Atlanta, GA  30366

  R.H. JONES JR.
  Ethyl Corporation
  P.O. Box 341
  Baton Rouge, LA  70821

  A. KAPOOR
  Buffalo Forge Company
  490 Broadway
  Buffalo, NY  14204

  Y.G. KARDISH
  Peabody Air Resources
  P.O. Box 5202
  Princeton, NJ  08540

  R.A. KENT
  Dart Industries Inc.
  P.O. Box 37
  Paramus, NJ  07652

  D. KERIVAN
  Cindaco Inc.
  2673 South Dixie
  Dayton, OH  45409

341

-------
B.N. KHETANI
Monsanto Envirochem
800 N. Lindbergh
St. Louis, MO 63166

P.A. KITTLE
Apollo Chemical Corp.
35 S. Jefferson Road
Whippany, NJ  07981

H. KROCKTA
The Ducon Co., Inc.
147 E. 2nd Street
Mineola, NY  11501

W.F. LALOR
Cotton, Inc.
4505 Creedmoor Rd., Box 30067
Raleigh, NC  27612

R.E. LANGLOIS
Owens Corning Fiberglas
Technical Center
Granville, OH  43023

J.W. LEATHERDALE
J.W. Leatherdale,  Inc.
382 Windsor  Street
N. Plainfield, NJ  07060

F. LeBLANC
Department of Biology
University of Ottawa
Ottawa, Ontario, Canada KIN 6N5

D. LEITH
Harvard Univ., Dept. E.H.S.
665 Huntington Avenue
Boston, MA   02115

E.D. LEMON
U.S. Borax Corporation
3075 Wilshire Blvd.
Los Angeles, CA  90010

B. LINSKY
West Virginia University
Department of Civil  Engineering
Morgantown,  WV   26506

H.P. LIU
U.S. Borax Corporation
Boron,  CA  93516
L.A. LOMBANA
BSP, Div. of Envirotech
One Davis Drive
Belmont, CA  94002

R.A. MACK
3M Company
P.O. Box 33331, Stop 71
St. Paul, MN  55133

V. MADDEN
Enviro-Test, Ltd.
8545 W. Col fax Avenue
Lakewood, CO  80215

J.D. McCAIN
Southern Research Institute
2000 9th Avenue South
Birmingham, AL  35205

R.W. McILVAINE
Mcllvaiue Company
2970 Maria Avenue
Northbrook, IL  60062

A. MEISEN
Dept. of Chemical Engineering
University of British Columbia
Vancouver, BC, Canada  V6T 1W5

B. MENNINGA
Griffin Pipe Products
2000 Spring Road
Oak Brook, IL  60521

C. MENOHER
American Air Filter
215 Central Avenue
Louisville, KY  40201

T. MORTON
Babcock £ Wilcox
609 N. Warren Avenue
Apollo, PA  15613

J.C. MYERS
Texas Air Control Board
8520 Shoal Creek
Austin, TX  78758

S.F. NEWTON
Tennessee Valley Authority
Wilhite Building
Muscle Shoals, AL    35660
                                        342

-------
P. J. NICOTRI
Environmental Elements Corporation
185 Arch Street
Ramsey, NJ  07446

G.M. OBELDOBEL
St. Joe Minerals Corp.
Route  18
Monaco, PA  15061

A.W. OBERHOFER
Nalco  Chemical Company
6216 W. 66th Place
Chicago,  IL  60638

R.G. OSTENDORF
Proctor £ Gamble
7162 Reading Road
Cincinnati, OH  45222

J-L.PAQUET
Aluminum  Co. of Canada
Arvida, Box 250
Quebec, Canada G7S 4K8

R.D. PARKER
Air  Pollution Technology,  Inc.
4901 Morena Blvd., Suite 402
San  Diego, CA  92117

L. PATTON
Florida Sugar Cane League
115  S. Lopez
Clewiston, FL  33440

B.O. PAUL
Stauffer  Chemical Company
P.O. Box  472
Mt.  Pleasant, TN  38474

R.G. PETING
Griffin Pipe Products
2000 Spring Road
Oak  Brook,  IL  60521

M.A. PETRILLI
Monsanto  Enviro-Chem
Box  14547 - Corp. Sq. Ofc.  Park
St.  Louis, MO  63178

M.J. PILAT
Dept.  Civil Engineering
Univ.  of  Washington
Seattle,  WA  98195
M.P.  POLAKOVIC
Bureau of Air Pollution Control
P.O.  Box 2807
Trenton, NJ  08625

S. POPKOFF
Research-Cottrell, Inc.
P.O. Box 750
Bound Brook, NJ  08805

N.E. PRANGE
Monsanto Enviro-Chem
Box 14547,  Corp. Sq. Ofc. Park
St. Louis,  MO  63178

D.A. PRICE
TRW - Energy Systems
One Space Park
Redondo Beach,  CA  90278

B. RAGLAND
Schneible Company
714 N. Saginaw Street
Holly, MI  48033

A.C. REEVE
Amoco Chemicals Corp.
P.O. Box 400
Naperville,  IL  60540

G. REY
U.S. Environmental Protection Agency-OEMI
Washington,  D.C.  20460

S.R. REZNEK
U.S. Environmental Protection Agency
Washington,  D.C.  20460
R. RHUDY
Bechtel Corporation
P.O. Box 3965
San Franciso, CA  94119

C.E. SCHAFER
Eastern Stainless Steel Co
P.O. Box 1975
Baltimore,  MD  21203

R.T. SCHNEIDER
Pridgen Engineering
P.O. Box 2008
Lakeland, FL  33803
                                        34.3

-------
G.C.  SCHREIBER
Schneible Company
P.O.  Box 100
Holly, MI  48442

G.L.  SCHROEDER JR.
Albert Switzer £ Assoc., Inc.
Suite 200, 3330 Lake Villa Dr.
Metairie, LA 70002

R.W.  SCHUFF
Fluid-Ionic Systems
2525  E.  Magnolia
Phoenix, AZ  85034

K.T.  SEMRAU
Stanford Research Institute
333 Ravenswood Avenue
Menlo Park, CA  94025

S.V.  SHEPPARD
Ceilcote Co., Inc.
140 Sheldon Road
Berea, OH  44130

S. SMALLWOOD
West  Virginia APC Commission
1911  Warwood Avenue
Wheeling, WV  26003

D.L.  SMITH
Conoco Coal Development
Research Division
Library, PA  15129

L.E.  SPARKS
I.E.R.L. - MD-61
U.S.  Environmental Protection Agency
Research Triangle Park, NC  27711

D.R.  SPINK
Dept.  Chemical Engineering
Univ.  of Waterloo
Waterloo, Ont., Canada N2L 3G1

B.A.  STEINER
Armco Steel Corporation
P.O.  Box 600
Middletown, OH  45043

D. STELMAN
Atomics International
8900 DeSoto
Canoga Park, CA  91304
J.G. STITES
UOP Inc., Air Correction Division
10 UOP Plaza
Des Plaines, IL  60016

H. SURATI
Beltran Associates, Inc.
1133 E. 35th Street
Brooklyn, NY  11210

D.O. SWENSON
Black § Veatch Consulting Engr.
1500 Meadow Lake Pkwy.
Kansas City, MO  64114

M.F. SZABO
Pedco Environmental, Inc.
11499 Chester Rd.
Cincinnati, OH  45241

D.R. TALBOT
Martin Marietta Corporation
1450 S. Rolling Road
Baltimore, MD  21227

A.M. TELFORD
Illinois EPA
1309 Dial Court
Springfield, IL  62704

K. WALKER
Environmental Elements Corporation
185 Arch Street
Ramsey, NJ  07446

J.G. WEIS
Burns § McDonnell Eng. Co.
4600 E. 63rd St. Trafficway
P.O. Box 173
Kansas City, MO  64141

E.D. WESSMAN
Dept. of Mechanical Engineering
Univ. of Waterloo
Waterloo, Ontario, Canada N2L 3G1

L. WHITAKER
Harrington  Industrial Plastics, Inc,
168 Freedom Avenue
Anaheim, CA  92801

F.A. WHITSON
TRW - Energy Systems
One Space Park
Redondo Beach, CA  90278
                                       344

-------
R. WIENER
Chemico Air Pollution Control
One Penn Plaza
New York NY  10001

D.A. WILSON
Aluminum Co. of Canada
Bldg. 270
Jonquieve, Quebec, Canada

G.J. WOFFINDEN
Meteorology Research, Inc.
464 W. Woodbury Road
Altadena, CA  91001

E.S. WYATT
Pridgen Engineering Co.
P.O. Box 2008
Lakeland, FL  33803

S. YUNG
Air Pollution Technology, Inc.
4901 Morena Blvd., Suite 402
San Diego, CA  92117

D.L. ZANDERS
Monsanto Research Corp.
123 Red Bud Court
Gary, NC  27511
                                       345

-------
                                 TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
1  REPORT NO.
  EPA-600/2-77-193
                                                       3. RECIPIENTS ACCESSION-NO.
4. TITLE AND SUBTITLE
 Second EPA Fine Particle Scrubber Symposium
            5. REPORT DATE
             September 1977
                                                       6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
 Richard Parker and Seymour Calvert
                                                       8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Air Pollution Technology, Inc.
 4901 Morena Boulevard, Suite 402
 San Diego, California 92117
            10. PROGRAM ELEMENT NO.
            1AB012; ROAP 21ADL-029
            11. CONTRACT/GRANT NO.

            68-02-2190
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC  27711
            13. TYPE OF REPORT AND PERIOD COVERED
            Proceedings; 9/76-7/77	
            14. SPONSORING AGENCY CODE
              EPA/600/13
^.SUPPLEMENTARY NOTES
 Mail Drop 61, S18/541-2925.
                              project officer for this report is Dennis C. Drehmel,
16. ABSTRACTTne repOrj- presents the proceedings, including introductory remarks and
 16 technical papers, of the Second  Fine Particle Scrubber Symposium, held May
 2-3, 1977, in New Orleans.  Sponsored by the U.S. Environmental Protection Agency,
 the symposium was  held to stimulate and generate new ideas for fine particle  control
 using wet scrubbers, and to promote the transfer of technology to users.  Subject
 matter concerned the collection of fine particles  in any type of wet collector with
 emphasis on scrubber performance data in industrial applications.
17.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                           b.lDENTIFIERS/OPEN ENDED TERMS
                         c. COSATI Field/Group
 Air Pollution
 Dust
 Scrubbers
 Industrial Processes
 Air Pollution Control
 Stationary Sources
 Particulate
13B
11G
07A
13H
13. DISTRIBUTION STATEMENT

 Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES

     362
20. SECURITY CLASS (Thispage)
Unclassified
                         22. PRICE
EPA Form 2220-1 (9-73)
                                         346

-------