EPA-600/2-76-133
May 1976
Environmental Protection Technology Series
                      EVALUATION OF  MAGNETICS FOR
                                 FINE  PARTICLE  CONTROL
                                     Industrial Environmental Research Laboratory
                                          Office of Research and Development
                                         U.S. Environmental Protection Agency
                                   Research Triangle Park, North Carolina 27711

-------
                RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped  into five  series. These five broad
 categories were established to facilitate further development and application of
 environmental technology. Elimination of traditional  grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The five series are:
     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.    Socioeconomic Environmental Studies
 This report has been  assigned  to the  ENVIRONMENTAL PROTECTION
 TECHNOLOGY series. This series describes research performed to develop and
 demonstrate instrumentation,  equipment, and methodology to repair or prevent
 environmental degradation from point and  non-point sources of pollution. This
 work provides the new  or improved technology required for the control and
 treatment of pollution sources to meet environmental quality standards.
                    EPA REVIEW NOTICE

This report has been reviewed by  the U.S.  Environmental
Protection Agency, and approved for publication.   Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service. Springfield, Virginia 22161.

-------
                                  EPA-600/2-76-133

                                  May 1976
              EVALUATION

            OF  MAGNETICS

   FOR  FINE PARTICLE  CONTROL
                     by

        K. P. Ananth and L.J. Shannon

          Midwest Research Institute
            425 Volker  Boulevard
         Kansas City, Missouri 64110
       Contract No. 68-02-1324, Task 26
            ROAPNo.  21ADL-029
         Program Element No. 1AB012


    EPA Task Officer:  Dennis C.  Drehmel

 Industrial Environmental Research Laboratory
   Office of Energy, Minerals , and Industry
      Research Triangle Park, NC  27711


                Prepared for

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

-------
                               CONTENTS

                                                                    Page

List of Figures	    iv

List of Tables	     V

Acknowledgments 	    VI

Summary	     1

Sections

I      Introduction  	     3

II     Particle Agglomeration in Magnetic Fields. .... 	     4

         Magnetic Fields and Charged Particles	     4
         Magnetic Fields and Uncharged Particles	    11

III    Magnetic Separation	    16

         Operation of a HGMS	    16
         Principle of a HGMS	    18
         Laboratory Study of HGMS	    24
         Comparison of Magnetic Force With Conventional
           Mechanisms	    24

IV     Conclusions	    26

References	    27
                                   iii

-------
                                FIGURES

No.                                                                Page

 1     The Magnetic Force  F  Acting on a Charged Particle.  ...     5

 2     Arrangement to Generate a Strongly Nonhomogeneous
         Magnetic Field H, Which Can Be Used to Separate
         Ferromagnetic Particles	    15

 3     Schematic Representation of "-im Type High Gradient
         Magnetic Separator 	    17

 4     Cross Section of Spherical Particle, Radius r_ Attached
         to Ferromagnetic Wire, Radius a, Magnetized by Uniform
         Magnetic Field HQ	    19

 5     Log-Log Plot of Magnetic (Fm) and Competing Gravitational
         (Fg) and Drag Forces  (F^) Versus Particle Size rp for
         Magnetized Wire Whose Size is Matched to That of Particle
         to be Trapped.  Computed for CuO Particle Attracted by
         Ferromagnetic Wire of Radius 3r_ Magnetized by 10 KOe
         Field Strength and Acted Upon by Slurry With Velocity
         of 5 cm/sec	    21

 6     Log-Log Plot of Magnetization Versus Particle Radii.  ...    23
                                    IV

-------
                                TABLES




No.                                                                Page





 1     Terminal Drift Velocity of Particles  in a Magnetic Field  .  .    7





 2     Maximum Particle Charge as a Function of Particle Size  ...   10

-------
                            ACKNOWLEDGMENTS

     The work presented in this report was performed by Midwest Research
Institute for the Industrial Environmental Research Laboratory-RTF of the
Environmental Protection Agency as Task Order No. 26 of Contract No.  68-
02-1324.  The work was performed by Dr. K. P. Ananth, Senior Environmental
Engineer, and Dr. L. J. Shannon, Assistant Director, Physical Sciences
Division.  The authors appreciate the helpful comments provided by
Dr. Dennis Drehmel, the EPA Project Officer, in conducting this study.
                                   VI

-------
                               SUMMARY

     Increasing concern for better control  of  fine  participates  has  di-
rected efforts toward development of new systems  that  are  inherently more
effective than conventional control devices or seeking systems that  can
improve the effectiveness of existing conventional  devices in capturing
fine particulates. This evaluation -.-as undertaken to determine the effect
of using magnetic forces to:  (a) agglomerate  fine  particulates  and  assist
their capture in conventional control devices} and  (b) directly  capture
fine particulates in magnetic devices.

     The use of magnetic fields to agglomerate either  charged or uncharged
particles is of no apparent utility in industrial gas  cleaning applica-
tions. In the case of charged particles, residence  times clearly in  excess
of those tolerable in industrial situations would be required to achieve
any appreciable agglomeration.

     Dipole forces, created by external magnetic fields, can influence
the agglomeration of uncharged particles. However,  significant enhance-
ment of agglomeration rates occurs only for ferromagnetic  materials  in
the presence of very strong magnetic fields. Since  the particulate pol-
lutants emitted from most industrial sources are not  ferromagnetic,  the
use of dipole forces caused by magnetic fields is of no real significance.
Also, it is quite likely that, even with ferromagnetic materials, unac-
ceptable residence times would be required to effectively  utilize dipole
forces.

     Direct capture of fine particulates should be readily achievable,
in principle, in magnetic separators provided the particles are  ferromag-
netic in nature and the magnetic force acting on them is greater than the
competing forces of gravitation and hydrodynamic drag. However,  as noted
previously, many particulates of interest are only weakly ferromagnetic
or paramagnetic in nature and high magnetic field gradients are needed
to achieve particle capture. Another potential problem could involve pro-
ducing the high gradients and large magnetic forces over a surface area
large enough to trap practical numbers of particles.

-------
     Currently, high magnetic field gradients have been generated  by  ap-
plying a relatively uniform background magnetic field to a ferromagnetic
structure (grids, screens, steel wool, etc.) and inducing magnetic poles
along properly oriented edges. A device using this approach is the high
gradient magnetic separator (HGMS) which has been successfully used in
mineral beneficiation and wastewater treatment. Despite the fact that
this device has not apparently been used for particulate collection in
industrial gas cleaning systems, it is attractive in principle, and war-
rants further investigation.

-------
                              SECTION I

                             INTRODUCTION

     The search for new systems to enhance fine particulate control  has
directed efforts toward evaluating the agglomeration of fine participates
by magnetic forces and toward evaluating magnetic devices  for direct cap-
ture of fine particulates. If magnetic forces are effective in agglomerat-
ing fine particulates, then the agglomerates can be easily handled  by
conventional control devices.

     Agglomeration of particles is a complicated process.  A variety of
forces causing the movement of particles toward one another may interact
to produce collisions and subsequent agglomeration. As the particles col-
lide and agglomerate, the concentration of each size changes with time,
each size interacts with every other size, and the agglomeration rates
change with size and concentration. Individual mechanisms  of particle
agglomeration will be more effective under certain conditions and the
optimum utilization of one or more mechanisms will require well designed
systems.

     Agglomeration of particles by magnetic forces is occasionally men-
tioned in the literature, but little effort has been devoted to either
theoretical or experimental studies of the process. Magnetic fields can
alter the motion of particles suspended in a gas stream depending upon
the magnetic permeability of the particles and magnetic fields might be
used to enhance the agglomeration of both charged and uncharged particles.
Particle agglomeration via the interaction of magnetic fields with charged
particles is reviewed in the next section. This discussion is followed
by discussions of magnetic fields and uncharged particles  and magnetic
separation, primarily the high gradient magnetic separation technique
(HGMS). A brief conclusion is also presented.

-------
                              SECTION II

              PARTICLE AGGLOMERATION IN MAGNETIC FIELDS

     Mechanisms by which particles may agglomerate in magnetic  fields
are reviewed in this section. The interactions of electrically  charged
particles with no intrinsic magnetic properties are discussed first fol-
lowed by a review of the influence of permanent and induced  dipole moments
on agglomeration.

MAGNETIC FIELDS AND CHARGED PARTICLES

     When an electrically charged particle with no intrinsic magnetic
properties moves in a magnetic field transverse to the  field lines» a
force is generated on the particle. The resulting force is  termed the
Lorentz force. If a particle carrying N elementary charges  "e"  moves
with a velocity vpj the direction of the resulting force will be at
right angles to both the direction of the field and the direction of
motion of the particle (see Figure 1). As a result of the Lorentz force,
the particle will be diverted from its original path. The magnitude of
this force is found to be directly proportional to the  velocity.  If the
velocity v  of the moving charged particle is not perpendicular to  the
direction of the magnetic field, but makes an angle 0 with the  field,
then the velocity vector vp may be resolved into two components:  Vp
cos 0 in the direction of the field, and vp sin 0 perpendicular to  the
field. In this general case, the force acting on the moving charged
particle is perpendicular to both the magnetic field and to vp  sin  0,
and has a magnitude proportional to vp sin 0, as shown in Figure 1.

     The potential for collisions and possible agglomeration of particles
can be qualitatively assessed by determining the terminal drift velocity
of a particle in a magnetic field. The terminal drift velocity  can be
obtained by equating the Lorentz force to the resistance of the gas,  cal-
culated from the Stokes-Cunningham Law:

-------
                             F =  VHsin
                                                 Direction of
                                                 Magnetic Field
                       Moving
                       Charged
Figure  1.   The magnetic force  F   acting on a charged particle,

-------
                                   3TT|J,gdvt
                           NqvpH =	                            (1)
where    N = number of charges on particle

         q — elementary charge

        Vp = velocity of particle in field

         H = magnetic field
    •••'.-
        Ug = gas viscosity

         d = particle diameter

        vt = terminal drift velocity of particle      '-

         C — Cunningham factor

 The left-hand term in Eq. (1) represents the Lorentz force and the right
 term the gas resistance. Eq. (1) can be rearranged to yield the following
 expression for  the terminal drift velocity:

                                NqvpHG
                           Vt = 3nn  d  '
 Eq.  (2)  indicates  that the  terminal drift velocity varies directly with
 the  number  of  charges on  a  particle, the particle velocity, and the mag-
 netic field and  inversely with the particle diameter.  Since the drift
 velocity varies  inversely with particle size, differentials in the drift
 velocity will  exist  in a  polydisperse  aerosol flowing  through a magnetic
 field. The  possibility of particle collision and agglomeration exists
 and  the  rate of  either process will be dependent upon  the magnitude of
 the  differential in  drift velocities and the residence time in the mag-
 netic field as well  as the  fluid mechanics of a specific system.

     Eq. (2) can be  used  to estimate values of the terminal drift velocity.
 Table 1  presents values of  vt calculated using typical values of the gas
 velocity encountered in gas cleaning operation. Gas  velocities of 5 and
 25 m/sec (1,000  and  5,000 fpm) were selected. Magnetic fields of 5 x 10^,
     and 10^ gauss were used (a  field  of  10^ gauss represents an efficient

-------
Table 1.  TERMINAL DRIFT VELOCITY OF PARTICLES IN A MAGNETIC FIELD

d
(cm) ,
1 x 10'4 v


1 x lO'4


I x ID'5


vp H
n (m/sec) (gauss)
1 - 5.1 5 x 103
100
1,000
i 104
100
1,000
1 . 105
100
1,000
1 25.5 5 x 103
100
1,000
1 104
100
1,000
1 105
100
1,000
1 5.1 5 x 103
100
1,000
1 10*
100
1,000
1 105
100
1,000
vt
(cm/sec)
2.4 x lO'7
2.4 x 10'5
2.4 x 10-4
4.8 x ID'7
4.8 x ID'5
4.8 x 10~4
4.8 x 10-6
4.8 x 10"4
4.8 x 10~3
1.2 x ID'6
1.2 x 10'4
1.2 x lO'3
2.4 x ID'6
2.4 x 10-4
2.4 x lO'3
2.4 x 10-5
2.4 x KT3
2.4 x ID'2
2.4 x 10-6
2.4 x ID'4
2.4 x 10~3
4.8 x 10-6
4.8 x 10"4
4.8 x lO'3
4.8 x lO-5
4.8 x 10~3
4.8 x ID'2

-------
Table 1.  (CONCLUDED)

d vp H ,
(rm> n (m/scc) (gauss)
v ' — - ' •
1 x 10-5 i 25.5 5 x 103
100
Jh* \f \S
1,000
1 104
L
100
^ ^/ x/ • 1^
1,000
i 105
1
100
1,000
vt
(cm/sec)
1.2 x 10"5
n
1.2 x KT3
1.2 x 10-2
2.4 x 10-5
2.4 x ID'3
2.4 x 10-2
2.4 x 10-4
2.4 x ID'2
i
2.4 x 10" L

-------
permanent magnet). Examination of Table 1 shows that the terminal drift
velocity is less than 0.01 cm/sec except at high particle velocities and
magnetic fields indicating that magnetic fields alone do not offer any
advantages as a mechanism for primary removal* of particles in a gas
stream.

      Since the force acting on the charged particle and the resultant
drift velocity are dependent upon particle size, the relative motion of
different-sized particles might be sufficient to bring them close enough
to each other to undergo collision and agglomeration. A detailed calcu-
lation of the path of particles in a magnetic field and determination of
whether or not collision and agglomeration would occur requires knowledge
of the magnetic field distribution, gas flow pattern, and charge distri-
bution on the particles.

      A qualitative assessment of the likelihood of collision and agglom-
eration can be obtained by considering the magnitude of the charge on
particles, the relative drift velocities between particles and the inter-
action of various particle collision mechanisms. At the outset it should
be noted that if  all the particles have a unipolar charge, collision is
unlikely because  particles of like charge will repel each other. If it
is possible to achieve bipolar charging, then collision and agglomeration
might occur. An estimate of the maximum number of charges that can reside
on a  particle can be obtained from various theoretical equations describ-
ing particle charging..!/ If only field charging is considered, Eq. (3)
can be used:
                           N    = 4TTe0rE0                            (3)
                            max      o p o
where      eQ = permittivity  of  free space

           rp = particle  radius

           Eo = applied electric field

Table  2 presents  a  tabulation of maximum particle charge  as  a function
of particle size  for  an  electric field of  2 x  106 volts/m.
*   i.e., a magnetic precipitator would not  be  feasible.

-------
                Table 2.   MAXIMUM  PARTICLE  CHARGE AS A
                      FUNCTION  OF  PARTICLE  SIZE
                         (E =  2 x 106  volts/m)
                    Particle  size
                          d                 Maximum number
                    (micrometer)           of charges,
                         100                  3.5 x 106
                          10                  3.5 x 104
                           1                     350
                           0.1                   3.5
     Inspection of Table 1 indicates that for the charge  levels  shown
in Table 2, relative terminal drift velocities for fine particles  of the
order of 10~" to 10~4 cm/sec result for reasonable values of Vp  and H.
Impaction mechanisms are precluded at such low relative velocities, but
interception and diffusion processes might result in collision and ag-
glomeration. Single particle target efficiencies of 0.2 to 0.5 for the
combined effects of interception and diffusion appear likely and some
collision and agglomeration may occur if the particles have sufficient
residence time in the magnetic field to permit the particles to  traverse
the average distance between particles (i.e., the interparticle  spacing
distance).

     The minimum residence time required for the particles to  traverse
the average distance between particles will depend upon the grain loading
of the aerosol. For aerosols with a grain loading ranging between 2.29
and 22.^9 g/m^ (i.e., 1 and 10 grains/scf), interparticle  spacing dis-
tances would be of the order of 10"^ cm. If the relative  drift velocities
are in the range of 10"4 to 10~6 cm/sec, residence times  of 10 to 1,000
sec would be required for particles to traverse a distance of  10"  cm.
A residence time of 10 sec might be feasible in an industrial  applica-
tion, but 1,000 sec is clearly not. A residence time of 10 sec would  re-
quire an effective magnetic field ranging from 45.72 to 228.6  m (i.e.,
150 to 750 ft) in length (depending on gas velocity). Thus, the  physical
size of equipment required to treat gas streams of 2,831.7 m3/min (i.e.,
100,000 cfm) even if cylindrical, spherical, or U-shaped systems were
designed, appears to be prohibitive.
                                   10

-------
     Even if agglomeration could be achieved with acceptable equipment
configurations, the use of magnetic fields to agglomerate charged parti-
cles appears to have dubious value. A particle charging system of some
sophistication would be required* as a component of the system as well
as a particle collection device downstream of the magnetic agglomeration
chamber. Since the particles will probably have residual charge,  an elec-
trostatic precipitator would seem to be a logical choice for the  collec-
tion device. If such is the case, it is difficult to justify placing a
magnetic agglomeration system between what would basically be an  electro-
static precipitator system.

     The preceding qualitative analysis results in the conclusion that
the use of magnetic fields to agglomerate charged particles is not attrac-
tive. Electrostatic agglomeration and subsequent collection in an ESP or
direct collection in an ESP are more viable approaches.

MAGNETIC FIELDS AND UNCHARGED PARTICLES

     Dipole forces, resulting from permanent dipole moments of aerosol
particles or dipole moments generated in a magnetic field, may be suffi-
cient to alter particle motion and cause collision and agglomeration.
Very little attention has been directed to the utilization of dipole
forces resulting from magnetic properties and fields for the agglomera-
tion of particles. The limited work on the subject is reviewed in the
next section.

Agglomeration of Particles with Permanent Dipole Moments

     Aggregates of particles may form in the absence of an external mag-
netic field if the particles have permanent dipole moments. On approach-
ing one another the dipoles orient themselves so that opposite dipoles
attract. Theoretical studies of the coagulation of magnetic spherical
dipoles have been conducted by Zebel.—  Zebel employed a statistical
mechanics approach in order to devise equations for the rate of coagula-
tion of weak and strong dipoles. For weak dipoles, Zebel derived  the fol-
lowing expression for the coagulation rate.
                        f =	1   ,   v                          (4)
                            1 - 0.0952/
   Bipolar charging would be necessary to avoid particle repulsion  which
     would occur between unipolar charged particles.
                                  11

-------
where      f =     coagulation rate with dipoles	
               coagulation rate in absence of dipoles
         r   = r  + r  = sum of radii of particles 1 and 2
          p  — dipole moment of particle 1

          p  = dipole moment  of particle 2

           k = Boltzmann constant

           T = absolute temperature

The term  r«  represents the distance of two dipoles,  with moments = pj
and  PT , at which their potential energy is equal  to  the thermal energy,
kT . Thus,  r^f/ri2 -^ 1  means that at the contact  distance  of  the parti-
cles their potential energy is large compared to the thermal energy; this
is a stro;-  interaction. Alternatively,  rM/r!2 <<: ^  represents a very
weak interaction. If  r^ « rio » Eq. (4) simplifies to:

                                      rr
                                       M
     For the case of strong dipole interaction,  the coagulation rate is
given by Eq. (6):
                                 M
                             f  = ~
                                 12
Eq. (6) is generally applicable when  r^ > 1.4

     Evidence of the experimental verification of either Eqs.  (4), (5),
or (6) was not found in the literature. References 3 and 4 discuss the
coagulation of particles which are permanent magnets. Iron and nickel
smokes, having particles which are permanent magnets, were generated by
heating the vapors of the carbonyls in the absence of oxygen.  Below the
Curie point (360° C for nickel and 770 C for iron) the smokes coagulate
                                   12

-------
to thread-like aggregates but above it rounded aggregates or very short
chains are formed.  Thread aggregates orient themselves in a magnetic field
and turn through 180 degrees when the field is reversed, thus showing that
the orientation is not dje to induction but due to permanent magnetism of
the particles.  How the magnetism originates is not known.  Smokes of iron
oxide crystals have similar properties but particles of amorphous iron ox-
ide are not ferromagnetic and do not give linear aggregates without an ex-
ternal field.

     It should be noted that very few particles possess permanent magnetic
properties (only iron, nickel, and cobalt and some alloys are ferromag-
netic).  Because only a few particles have permanent magnetic properties,
limited opportunities will exist to exploit particle agglomeration by this
technique.  Agglomeration by this mechanism will occur naturally whenever
it is possible and efforts to enhance the process appear to have little
merit.

Dipole Moments Generated in a Magnetic Field

     Dipole moments can be generated when uncharged particles are placed
in a magnetic field.  If a homogeneous magnetic field of strength  H  is
used, it will polarize spherical aerosol particles with radius  rp  to
dipoles having moments:
                                     •*  f\
                                                                       (7)


where     u = magnetic permeability of particle.

     By making simplifying assumptions regarding particle motion and assum-

ing that  V1      is equal to unity (i.e., ferromagnetic material), Fuchs^-'
          u + 2
has shown that a distinct increase occurs in the factor  f  (see Eq. (4))
when rM/rJ -> 20.

     Under the above conditions (i.e., strong attraction), the ratio of
the rate of coagulation with and without dipole forces is:
                                    23
                                                                       (8)
From the preceding it is readily apparent that acceleration of agglomera-
tion of fine particles by induced magnetic dipoles will only occur in
strong fields aid for ferromagnetic materials.
                                    13

-------
     If a nonhomogeneous magnetic field (Figure 2) is used, a force will
act on the particles which might result in deflection of particles with
resultant collision and agglomeration. For this case, Zebel presents the
following equation for the
                           K = a/2 grad H                            (9)
               n. _ i  ^
where      a =
           u- = magnetic permeability of particle

          r  = particle radius
           P

Detailed  calculations  of  the path of particles and determination of
whether or not  they will  collide and agglomerate require a knowledge of
the magnetic  field distribution, particle size distribution, particle
grain  loadings, magnetic  permeabilities of particles, and the gas flow
pattern.  As was noted  previously, only few materials possess high mag-
netic  permeabilities and  it is  likely that only ferromagnetic materials
would  show any  significant agglomeration.
                                   14

-------
Figure 2.  Arrangement to generate a strongly
  nonhomogeneous magnetic field H, which can
  be used to separate ferromagnetic particles
  (N and S designate the north and south pole,
  respectively).
                     15

-------
                             SECTION III

                         MAGNETIC SEPARATION

     Magnetic separation is an old technique whose use has been restricted
in the past to the separation of strongly magnetic materials like iron and
magnetite. A good example of this type of separator is the conventional
drum-type magnetic separator.—'  Currently, magnetic separation of weakly
magnetic materials (weak paramagnetic materials) is made possible with
the aid of high magnetic field gradients. One such device is the "high
gradient magnetic separator" (HGMS). Since particulates emitted from most
industrial sources are not ferromagnetic, this discussion will be limited
to the more widely applicable HGMS.

     The literature contains several articles on HGMS and its operating
principle and potential application areas.	   The application areas
discussed include mineral beneficiation (of taconite iron ores), coal
desulfurization, and wastewater treatment. Unfortunately, none of the
papers discuss the potential use of HGMS in controlling particulates in
industrial gas streams. However, the principle and operation of the de-
vice are independent of the application area. A brief discussion of the
HGMS concept is presented next.

OPERATION OF A HGMS
                                                 7,9/
     Figure 3 is a schematic of a Kolm type HGMS."""^  A magnet designed
to produce a strong variable field in the canister volume is used. This
volume is packed with  a matrix of filamentary, ferromagnetic material.
These filaments produce large surface areas of high magnetic gradients
along their edges and  they resist compression and clumping in the applied
magnetic field. The filaments are chosen to match the size of the parti-
cles to be removed in  order to optimize the magnetic forces. The feed    ,
in a fluid (usually water) slurry, is passed down through the canister.""
The fluid and nonmagnetic particles pass easily through the relatively
open structure of the matrix. The trapped magnetic particles are easily
washed out when the applied field is reduced to zero.-^
                                  16

-------
              Magnet
              Field
              Stainless
              Steel Wool
              Matrix
             Filtered^.
             Liquid Out
Figure 3«   Schematic representation of Kolm  type high gradient
                     magnetic separator.£jZ'
                                 17

-------
PRINCIPLE OF A  HGMS

     The principle of  operation of a HGMS  is  reviewed  in detail in Ref-
erence  7. A brief summary  of  the  review highlighting those aspects which
are  critical to particulate collection in  industrial gas streams is pre-
sented  in the  following  paragraphs.

     A  simple  expression for  the  magnetic  force acting on a particle is
 (in  the x component):
                        Fx  = / (Mp  - Mm) 5it • dV
                            J          dx

 In  Eq.  (10),   M^   and M  are  the  magnetization of the particle and medium,
 respectively;  V   is the  volume  of the particle and  dH/dx  the magnetic
 field gradient. According to  Eq,  (10), for a specific particle size (volume)
 any increase  in magnetization  or the  field gradient should improve the mag-
 netic force.  The magnetization (M) itself  is a function of the susceptibil-
 ity* and  the  magnetic field strength  as shown in Eq. (11).

                                 M  = xB                                 (11)

 Substitution  of Eq.  (11)  in Eq.  (10)  and integration yields;
                    Fx =  (v   -  v  )   .  .
                      x     P    ™         dx

where  b   is  a point  within the particle having volume  V   (see Figure 4).

     In Eq.  (12),   Yp,  Xjn  an<^  V  are intrinsic properties. The only
parameters  that  can be varied to  enhance  FX  are the magnetic field
strength   B(b)   and the field gradient  dH(b)/dx.  Even with a constant
field strength,  the magnetic  force could be improved by altering the
field gradient.  For an applied  field  HQ  less than bulk satuation  field
H  ,  along  the axis,  the  magnetic field gradient at the particle is given;—

                              JU       Q 2
                              v^n _ _ Ou  o
                              dx     ° b3

Therefore,  there is a certain filament radius (a) which will maximize  the
magnetic force on a given particle.  Oberteuffer-i' reports  that this
   Magnetic susceptibility,  y = u/uo -  1  where "u"  is  the magnetic  per-
     meability of the material (particle) and "uo" is the magnetic permea-
     bility of vacuum.
                                   18

-------
                                              FERROMAGNETIC
                                                  WIRE
Figure 4.  Cross section of spherical particle,, radius  r
                                                       ^
  attached to ferromagnetic wire,  radius a, magnetized
               by uniform magnetic field H .
                            19

-------
radius should be of the same order of magnitude as the particle radius and,
to be more precise, is three times the particle radius.  One apparent dis-
advantage is the fact that a high efficiency of magnetic separation can be
achieved only for certain sized particles in a polydispersed gas stream.

     It is also to be noted that if the magnetic force falls off signifi-
cantly over a distance of one particle diameter, each trapping site in
the separator will have a capacity of approximately one particle.  There-
fore to increase the number of sites in separators it is desirable to use
fibers or wires since the number of trapping sites is approximately the
same as the volume of the fibers.!/

     Another consideration in the HGMS is the need to determine the range
of particle radius over which the magnetic force is greater than the com-
peting forces of gravity and hydrodynamic drag.  The gravitational force
is given by:

                        Fg = ^ Tir^  (pp - pf) g                        (13)

where  r   is the particle radius,  p   and pr  the densities of the par-
ticle and medium, respectively, and  g  the local accleration of gravity.
Since  F   is a function of  r^   the gravitational force will be signi-
ficant for large particles.  The hydrodynamic drag force given by:

                              Fd =  6T77]rpv                             (14)

is significant  for small particles.  In Eq. (14),  T]  is the viscosity of
the medium and  v  is the velocity  of the particle relative to the stream.
Figure 5 shows  the magnetic and competing forces versus particle size for
a gradient-matched separator where  the filament size is matched with the
size of the particle to be trapped.—'

     It can be  observed from Figure 5 that the magnetic force (F ) exceeds
the hydrodynamic drag force (Fd) in a water slurry, only for particles of
size greater than 1 to 10 urn.  This result means that for CuO particles
< 1 urn in size  the hydrodynamic drag force predominates in a water slurry
system and capture by magnetic forces is ruled out.  We have extended this
analysis to an  air system to determine the applicability of magnetics in
gas cleaning which is of greater concern to us.  The Fd lines for air at
two different velocities have been  calculated and included in Figure 5.
It is interesting to note that in an air system the magnetic force (Fm)
dominates the drag force (Fd) even  for particles < 0.1 urn at a velocity
of 5 cm/sec.  Higher particle velocities shift the effectiveness of the
magnetic force toward larger particles.  For an air velocity of  50 cm/
sec,  which is close to the typical value for an industrial gas system,
the magnetic force is effective down to about 0.3 urn.

                                    20

-------
   X
  -o
       1Q
 ,-2
       10'
         ,-4
       10
 ,-6
       10'
         r8
      10'
rlO
                 Fd (Water) @ Velocity  of 5 cm/sec


               Fd (Air) @ Velocity of 50 cm/sec
                             Fd (Air) @ Velocity of 5 cm/sec
                    J_
10
                      -   O.lcm .
                      _ I     I
                                          10
          0.01 0.1   1       100        1       100

                      PARTICLE RADIUS, rp (/Am )
Figure 5.  Log-log plot of magnetic (Fm) and competing gravitational
  (Fg) and drag forces (F^) versus particle size r_ for magnetized
  wire whose size is matched to that of particle to be trapped. Com-
  puted for CuO particle attracted by ferromagnetic wire of radius
  3r  magnetized by 10 KOe field strength and acted upon by slurry
  with velocity of 5 cm/sec.2/
                                  21

-------
     If, instead of CuO particles, one uses Fe30^ which has a much higher
magnetic permeability, then the line depicting magnetic force (Fm) in
Figure 5 would move to the left.  This shift is shown by the dashed line.
The shift in the line caused by particles of high magnetic permeability
would thus enable the capture of particles well into the submicron region.

     For a given range of magnetization, the range of particles separable
in a HGMS and in an "ideal" separator have also been reported for a water
slurry system.^'  The "ideal" separator has a wire (filament) radius ap-
proximating the particle size (i.e., it is gradient-matched) and is as-
sumed to operate at a magnetic field intensity of 10 KOe* and a water
slurry velocity of 5 cm/sec.  The lower and upper limits of particles that
can be captured in such a device have been calculated by equating the com-
peting forces, i.e., magnetic force (Fm) with the drag force (F
-------
    103
     102
      10

c
at
o

c
s

8
o>
    10
      -1
    10
     -2
        0.01                  0.10                    1.0


                                  Particle  Radius,fp (fj.m )



         Figure 6.  Log-log plot  of  magnetization versus particle  radii
10
                                      23

-------
     *  Gas stream  flow rates/velocities.

     *  Gas stream  viscosity and density.

     *  Selection of appropriate wire diameter.

     *  Selection and generation of required field intensities.

LABORATORY STUDY OF HGMS

     Unfortunately, there is no direct experience available in the use of
such a  system for particulate  control in industrial gas streams.  Appar-
ently,  the only study that  is  being conducted to determine the use of a
HGMS in scrubbing fine particles in gas streams is a laboratory study at
Research  Triangle Institute (RTI) under EPA sponsorship.  The gas stream
under test at RTI is dust from the basic oxygen furnace in prefiltered
room air.  Efficiencies of  the order of 90% for particulates > 0.3 um have
been observed with  an optical  particle counter in the preliminary experi-
ments —  and  no other data  are available.

COMPARISON OF MAGNETIC FORCE WITH CONVENTIONAL MECHANISMS

     It is interesting to compare particulate capture due to conventional
mechanisms such as  impaction,  interception, and diffusion with that due to
magnetic  forces. Let us take a case where there is no magnetic force acting
on  the  particle and the wires  act as targets for interception, inertial im-
paction,  and  diffusion of the  particles at a gas velocity of 5 cm/sec.  Let
us  also assume that the wire diameter is three times the particle diameter
in  order  to conform to a gradient -matched magnetic separator.

     The  interception efficiency is given by:—'

                         T(  =  1 + R - - 1 -                          (15)
                                      (1 + R)
where   R  = ratio of particle  diameter to that of the intercepting body
(in our case,  R = 0.33).  The  inertial impaction parameter is given by:—'

                                        2
                             .   C pp v Dp
                                - ~                                   (16)
                                 18ugDb104

where     C = empirical correction  factor for resistance of gas  to move-
              ment of small particles

            = 1  (assumed  for  2 um particle)
                                    24

-------
         p  = particle  density in g/cc = 2.8 g/cc (equivalent to iron
              foundry particles)

          v = velocity  of particle in cm/ sec

            = velocity  of gas,  =  5 cm/ sec

         D  = particle  s i . ;-c in urn
         Ug  = viscosity  of  gas  in poises =-- 1.8 x 10   poises (for air)

         D,  = diameter of wire  in um

             = 6  um

     The diffusional  efficiency is;-!-t'

                    1], = H_ (1  + 0.55 Re1/2 Sc1/3)                     (17)
                          Pe  n

                                       n v
where    Pe  = peclet  number = Re Sc - —I —
                                        £
                                 I) vp
         Re1  = Reynolds number = — ^ -
                                  I' g

         Sc  - Schmidt number =  li&
                                p^

In  Eq .  (17),  I)    is  the particle diameter (2 urn) ,   v  is the particle ve-
locity  (5  cm/sec),  P   the density of the particle (2.8 g/cc),  ug - vis-
cosity  of  air and  S  the diffusivity of t he particle (assumed to he  2.7  x
     For  the'  conditions  indicated  above the intercept' ion efficiency  is cal-
culated lo be r)87.  and  the  impact ion parameter is 0.','9.   At the  latter value
the impact ion efficiency  is  x.ero. — -'   The diffusional efficiency  is  • 17,.—'
If, under the above  conditions,  magnetic forces can help capture  all of  the
2-\nn particles,  I hen systems ntili/.ing mai'.netic forces could be  rated as
being superior  to  those  de-vices  which only make use of the conventional
mechanisms.   Inspection  of Figure  5 shows that  at a gas (air) velocity of
5 cm/sec  the  magnetic  loree  is  effective up to about 0.05 |im.  One obvious
route- to  enirincr the performance- ol  a magnetic separator is to use magnetic
seeding e> I the-  gas stream  to he-  scrubbed with high penm-ab i 1 i t v  p.irticu-
late-s.   However, on  an  industrial  basis this approach does not appear via-
ble-.

-------
                              SECTION IV

                              CONCLUSIONS

     The use of magnetic fields to agglomerate either charged or uncharged
particles is of no apparent utility in industrial gas. cleaning applica-
tions. In the case of charged particles, residence times clearly in excess
of those tolerable in industrial situations would be required to achieve
any appreciable agglomeration. Only ferromagnetic materials could be ag-
glomerated by dipole forces and the paucity of ferromagnetic materials
in most industrial emissions makes that approach nearly useless.  Also,
it is quite likely that unacceptable residence times would be required
to effectively utilize dipole forces.

     Direct magnetic separation on the other hand appears to be of poten-
tial value for fine particulate control in industrial gas streams. However,
it has never been used for this application and only a laboratory scale
study is currently under way to test the HGMS for particulate control in
specific gas streams (i.e., streams containing high permeability particles)
A detailed review of the results of the present -study including aspects
such as fractional efficiency versus magnetic permeability of the particle,
particulate entrainment, wire (filament) diameter, and pressure drop would
be critical prior to further investigation. Also, the economic aspects of
the device, particularly the energy required to generate large magnetic
gradients in field size units, should be evaluated in the absence of super-
conducting magnets. Thus, it can only be said that the HGMS is attractive,
in principle, for particulate control in gas streams and warrants further
investigation to assess its actual utility.
                                  26

-------
                              REFERENCES

 10  Robinson, M., Electrostatic Precipitation, published in Air Pollution
       Control, Part I, Wiley-Interscience (1971).

 2.  Zebel, G., Staub. 19(11), 381 (1959).

 3.  Berscher, D., and A. Winkel, A. Phys. Chem., H6A, 1 (1936).

 4.  Berscher, D., and A. Winkel, Naturwiss., 2J5, 420 (1937).

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

 6.  Zebel, G., Staub (English Iraslation), 28(7), 1 (1968).

 7.  Oberteuffer, J. A., IEEE Transactions on Magnetics, MAG;40(2), 223,
       June 1974.

 8.  Kolm, H., J. Oberteuffer, and D. Kelland, Scientific American, p.
       47, November 1975.

 9.  Oberteuffer, J. A., IEEE Transactions on Magnetics. MAG-9_(3), 303,
       September 1973.

10.  Kelland, D. R., IEEE Transactions on Magnetics, MAG-9(3), 3079
       September 1973.

11.  Trindade, S. C., and H. H. Kolm, IEEE Transactions on Magnetics,
       MAG-9(3), 310, September 1973.

12.  De Latour, C., IEEE Transactions on Magnetics, MAg-9^3), 314, September
       1973.

13.  Private communication with Mr. Charles Gooding, Research Triangle
       Institute, North Carolina, February 9, 1976.

14.  Strauss,  W., Industrial Gas Cleaning. Pergamon Press, New York (1966).

                                     27

-------
 1. REPORT NO.
                                  TECHNICAL REPORT DATA
                           (Please read Inunctions on the reverse before completing)
  EPA-600/2-76-133
 I. TITLE AND SUBTITLE
                             2.
                                                        3. RECIPIENT'S ACCESSION-NO.
 Evaluation of Magnetics for Fine Particle Control
              5. REPORT DATE
               May 1976
                                                        6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
              8. PERFORMING ORGANIZATION REPORT NO.
 K. P. Ananth and L. J.  Shannon
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Midwest Research Institute
 425 Volker Boulevard
 Kansas City, Missouri  64110
              1O. PROGRAM ELEMENT NO.
              1AB012; ROAP 21ADL-029
              11. CONTRACT/GRANT NO.

              68-02-1324,  Task 26
 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
              Task Final; 2/75-2/76	
              14. SPONSORING AGENCY CODE
               EPA-ORD
 15. SUPPLEMENTARY NOTES Task officer for this report is D. C.  Drehmel, Mail Drop 61,
 Ext 2925.
 16. ABSTRACT
 The report gives results of an evaluation of the effectiveness of magnetic
 agglomeration/separation techniques in enhancing fine particulate capture.  Whereas
 residence times and magnetic fields required to achieve agglomeration via
 magnetic forces appear to be unrealistic, magnetic separation--especially high
 gradient magnetic separation--appears to be attractive in principle.
 7.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
  b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
 Air Pollution
 Magnetic Properties
 Agglomeration
 Separation
 Dust
  Air Pollution Control
  Stationary Sources
  Fine Particulate
 13B
 20C
                            11G
 3. DISTRIBUTION STATEMENT

 Unlimited
  19. SECURITY CLASS {ThisReport)
  Unclassified
21. NO. OF PAGES
    34
  20. SECURITY CLASS (Thispage)
  Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
28

-------