Anal Bioanal Chem (2004) 378 : 16931699

DOI 10.1007/s00216-003-2354-7

REVIEW

Hitoshi Watarai Masayori Suwa Yoshinori Iiguni

Magnetophoresis and electromagnetophoresis of microparticles in liquids

Received: 4 August 2003 / Revised: 16 October 2003 / Accepted: 17 October 2003 / Published online: 13 December 2003
Springer-Verlag 2003

Abstract The magnetic field-induced migration of particles in liquids is a highly-promising technique for the micro-separation analysis of bioparticles, such as cells and
large DNA. Here, new methods that make use of magnetophoresis and electromagnetophoresis to induce the migration of microparticles in liquids are briefly reviewed.
Magnetic force and Lorentz force are utilized in the new
methods. Some typical examples of the use of these methods are described, and the advantages of using a superconducting magnet for them are demonstrated.

ing new migration analysis techniques for single bio-microparticles. In this review, our recent studies on the magnetophoresis and electromagnetophoresis of microparticles
in liquids are summarized and their advantages are demonstrated.

Keywords Magnetophoresis Electromagnetophoresis

Bioparticles Superconducting magnet

Under a magnetic field, B, the magnetic potential energy

of a particle with volume V and volume magnetic susceptibility p can be written as follows:

Introduction

8 =

Most molecules in nature exhibit their functions by forming aggregates or microparticles. Living cells, DNA and
proteins all belong to this category. Indeed, in the case of
bioparticles, every particle may have an individual property that is slightly different from the properties of the
other particles. However, ordinary separation methods are
unable to analyze such individual particles; although many
methods for the separation or migration analyses of small
ions and small molecules have been developed (as typified by the current popularity of HPLC and CE), for microparticles of large molecular weight, including DNA
greater than 40 kbp and proteins larger than 104 Da, only
limited methods (such as sedimentation and cell sorting)
are currently available. Therefore, we need to invent new
analytical principles that can separate and characterize individual bioparticles in liquids.
Magnetophoresis and electromagnetophoresis, which
both use a magnetic field to cause migration, are promis-

where 0 is the vacuum magnetic permeability, B is the

magnetic flux density, and m is the volume magnetic susceptibility of the medium. A magnetic force working on a
particle can be written as

H. Watarai () M. Suwa Y. Iiguni

Department of Chemistry, Graduate School of Science,
Osaka University, Toyonaka, 560-0043 Osaka, Japan
e-mail: [email protected]

Magnetophoresis
General principles of magnetophoresis

S P 


) = JUDG8 =

9% 

(1)

S P 

9 %  %
(2)

Therefore, an inhomogeneous magnetic field can exert a
force on a particle in liquid. During the migration of the
particle in liquid, the drag force (as expressed by Stokes
Law) acts on the particle, FD = 6rv, where is viscosity and r the radius of the particle. Therefore, the magnetophoretic velocity can be expressed as follows:
Y=

 S P  
U  %  %



(3)

Therefore, the magnetophoretic velocity is directly proportional to (pm), r2, and (B)B.
Analytical separation techniques
Field-Flow Fractionation (FFF) is an effective method for
the separation of microparticles in liquid [1]. An external

1694

field is applied perpendicularly to the flow of the medium

in the FFF channel. The perpendicular position of a particle in the FFF channel depends on its physical properties.
Therefore, the elution time of each particle is determined
by the laminar flow distribution of the medium. Magnetic
FFF was first demonstrated by Vickrey and Garcia-Ramirez
[2], and Schunk et al. used a more conventional FFF channel and an electromagnetic solenoid coil [3].
Split-Flow Thin (SPLITT) fractionation is a close relative of FFF since it applies various external fields and has
the same direction of driving force, perpendicular to the
flow [4, 5]. The advantage of this method is its ability to
continuously separate microparticles. A magnetic field was
utilized as a driving force for SPLITT fractionation by
Fuh and co-workers [6, 7, 8]. This technique allowed the
complete separation of paramagnetic ion-labeled particles
(yeast, red blood cells, silica particles, and so on) from

Fig. 1 Schematic drawing of the configuration of the magnets and

the pole pieces in the capillary magnetophoresis

Fig. 2 Photographic representation of the magnetophoretic trapping of red blood cells at a flow rate of 1.0 L h-1. The magnetic
force acted on the red blood cells to push them downward, and the
medium that contained 0.1 M MnCl2 moved upward. Therefore,
the red blood cells were trapped at the position where the magnetic
force balanced with the drag force due to the bulk flow. Pictures
are shown at 3 s intervals from panels 1 to 6

non-labeled particles [6]. The magnetic susceptibility of

the microparticle could be determined to a good precision
by the fractional retrieval [8]. Zborowski et al. constructed
a cylindrically-symmetric SPLITT fractionation channel
that used a quadrupole magnetic field [9, 10, 11, 12], and
using this immunomagnetically-labeled human white blood
cells could be separated from non-labeled cells [12]. The
throughput of 107 cells/s of their system was much higher
than that of fluorescence activated cell sorting (103 cells/s),
which is the most widely-used technique applied to cell
analysis and sorting.
The methods mentioned above are similar to each other
because the direction of magnetic force is perpendicular
to that of the carrier flow. Watarai and Namba demonstrated the trapping of polystyrene latex particles and red
blood cells via magnetic force whose direction was parallel and opposite to that of the bulk flow [13, 14]. The configuration of the magnets (Nd-Fe-B) and iron pole pieces,
shown in Fig. 1, generated high magnetic field gradient in the
direction of the flow. The maximum value of (B)B/2,
of about 2000 T2 m1, is the highest magnetic field gradient generated with such a small permanent magnet, to our
knowledge. In their system, the difference in magnetic
susceptibility between the particle and medium, (pm),
was negative, because the aqueous medium contained a
paramagnetic ion (Mn2+). Therefore, the magnetic force
acted on the particle along the direction that the magnetic
field became weaker, as predicted by Eq. 2. As the bulk
flow was in the opposite direction to the direction that the
field became stronger, the particles were trapped at the
position where the drag force by bulk flow was balanced
by the magnetic force. Figure 2 indicates the magnetophoretic trapping of red blood cells. Changing the flow rate
allowed us to control the trapping efficiency. Figure 3
shows the separation of three polystyrene particles of different size by this method [13]. Using this technique, the
microparticles can be manipulated like using optical
tweezers.

Fig. 3 Counter current magnetophoretic fractionation of polystyrene particles with diameters of: a 2.77; b 5.87; c 9.14 m in
0.6 M manganese(II) chloride solution

1695

Single microparticle analyses

by magnetophoretic velocimetry
The magnetophoretic velocity is proportional to the volume magnetic susceptibility of the microparticle, as indicated by Eq. 3. Therefore, the velocity measurement can
determine the magnetic susceptibility and more information can be obtained from this magnetic susceptibility, as
discussed below.
Particle Tracking Velocimetry (PTV) and magnetophoresis of immunomagnetically labeled cells were combined
by Zborowski et al. [15, 16, 17, 18, 19]. They measured
the magnetophoretic velocity of an individual labeled cell
under an inhomogeneous magnetic field, and the magnetophoretic mobility was determined. In their theoretical
treatment, the magnetophoretic mobility depended on the

Fig. 4 An example of magnetophoretic behavior of MnCl2 aqueous droplets dispersed in 2-fluorotoluene phase. The concentration
of Mn(II) was 0.04 M

Fig. 5 The magnetophoretic mobility as a function of the concentration of MnCl2 in the droplets. The solid line was the calculated
value using Eq. 3

number of magnetic labels. The immunomagnetic label,

which attached to the antibody, bound to the cell by antigen-antibody reaction. The number of labels was proportional to the number of antigens on the surface of the cell.
Therefore, the number of antigens on the cell surface could
be determined from magnetophoretic mobility. Antibody
Binding Capacity (ABC) was defined as the quantity representing how many magnetic nanoparticles could bind
per single cell, and the difference in ABC between individual cells was investigated [19].
The magnetophoretic velocity of the macrophagic cell
that was fed the magnetic nanoparticle was measured, and
the quantity of the uptaken nanoparticles was demonstrated
by Wilhelm and co-workers [20]. The number of magnetic
particles in the cell, as determined by magnetophoresis
and ferromagnetic resonance, were compared. The ferromagnetic resonance technique could measure only the average value of the magnetic particles, but magnetophoresis could distinguish the differences between individual
cells. The average number as determined by magnetophoresis was in good agreement with the corresponding
ferromagnetic resonance measurement.
The determination of paramagnetic ions in a single
droplet using magnetophoretic velocity was demonstrated
by Suwa and Watarai [21]. The magnetophoresis of manganese(II) aqueous microdroplets dispersed in an organic
medium was observed and the migration velocity was
measured. The magnetic susceptibility was calculated using Eq. 3. The magnetic susceptibility of a solution can be
calculated easily by the sum of the product of concentration and molar magnetic susceptibility for each solute and
solvent. Therefore, the concentration of manganese(II)
could be determined from the magnetophoretic velocity.
We call this principle Magnetophoretic Velocimetry. When
using the magnetic field gradient generated by a pair of
permanent magnets, the lower detection limit of manganese (II) was 1016 mol/droplet.
Recently, a superconducting high field magnet has become available in our laboratory and magnetophoretic velocimetry has been carried out under the high magnetic
field gradient generated by it (10 T) [22]. The magnetophoretic capillary cell was set between a pair of iron pole
pieces, positioned in the bore of the superconducting magnet, where the magnetic field was the strongest and most
homogeneous. The value of (B)B/2 around the pole
pieces was determined from the magnetophoretic velocity
of polystyrene microparticles dispersed in manganese(II)
aqueous solution, and the maximum value was determined to be 4.7104 T2 m1, which was almost 100 times
larger than that attained by using a small permanent magnet (0.4 T). Under the high magnetic field gradient, the
magnetophoretic behavior of a single aqueous droplet containing MnCl2 was measured as shown in Fig. 4. The magnetophoretic velocity at the edge of the pole pieces was
plotted against the concentration of MnCl2, as shown in
Fig. 5. The velocity was divided by the square of the radius in order to normalize by the size. The magnetophoretic velocity was proportional to the concentration of
Mn(II). The lower detection limit of manganese (II) was

1696

Fig. 6 The plot of the magnetic susceptibility determined from the

magnetophoretic velocity versus the reciprocal of the radius

~1017 mol/droplet in this system. It should be emphasized

here that the magnetophoretic velocity could be observed,
even though the volume magnetic susceptibilities of both
medium and droplet were negative, under the high magnetic field gradient.
Furthermore, the formation of paramagnetic complex,
Dysprosium(III) laurate, at the interface of a single droplet
has been detected recently [23]. The magnetophoretic velocity of a 2-fluorotoluene droplet including lauric acid was
observed in an aqueous solution of Dy(III) ion. The magnetic susceptibility of the droplet, as calculated from the velocity, was proportional to the inverse of the size, as shown
in Fig. 6, and this clearly suggested the interfacial adsorption of Dy(III) complex. The interfacial concentration was
estimated as 2.61010 mol cm2. Therefore, it was demonstrated that the magnetophoresis is highly useful not only
for the separation of microparticles, but also for the detection of the adsorbed species at the liquid/liquid interface.

Electromagnetophoresis
Principles of electromagnetophoresis
The general concept of electromagnetophoresis (EMP) is
the application of Lorentz force to migration analysis. Electromagnetophoresis is a phenomenon where particles in an
electrolyte solution migrate in the direction perpendicular to
both a magnetic field and an electric current when the electric current is applied through the conductive fluid and the
homogeneous magnetic field is perpendicular to the current.
The basic theory of electromagnetophoresis was proposed by Kolin in 1953 [24]. When an electric current of
density j (A m2) passes through a conductive fluid of volume V (m3) perpendicular to a magnetic field, the fluid invariably experiences a force, the Lorentz force, expressed
as
) =  +M9

(4)

Fig. 7 Schematic drawing of the electromagnetic weight, FEMW,

and the electromagnetic buoyancy, FEMB, exerted on a particle.
They are perpendicular to a homogeneous magnetic field (B) and
to an electric current (i). The magnetic field is maintained at right
angles to the current

where 0 is the vacuum magnetic permeability (N s2 C2),

is the magnetic permeability of the conductor, and H is
the magnetic field strength (A m1). As the current passes
through the conductive fluid under a magnetic field, all of
the fluid experiences the force and so a pressure gradient
is produced inside. Figure 7 shows a schematic drawing of
forces exerted on a particle in the fluid in this situation.
When a current is applied through the enclosed conductive fluid in the homogeneous magnetic field, the force is
exerted on both the fluid and the particles. If the force exerted on the fluid is equal to the one acting upon the particles, the particles dispersed in the fluid will be at rest.
However, if the two forces are different, migration of the
particles will be caused. The two forces acting on a particle in Fig. 7 are called the electromagnetic weight (EMW),
FEMW, and the electromagnetic buoyancy (EMB), FEMB,
acting opposite to FEMW. The strength of FEMB is equivalent to the force working on the fluid whose volume is
equal to that of the particle (V).
The electromagnetophoretic force, FEMP, exerted on a
particle in a closed cell which has an inner section of S
(m2) is expressed as the sum of FEMW and FEMB:
S I L
)(03 = )(0: + )(0% =  %9

 I + S 6

(5)

where B is the magnetic flux density (N A-1 m1), p is the

electrial conductivity (S m1) of the particle, f is the electrical conductivity (S m1) of the medium, and i is the current (A). The electromagnetophoretic velocity vEMP of a
spherical particle in the fluid is expressed as
Y(03 =

 S I

  I + S

L%U 

6 &:

(6)

1697

where is the fluid viscosity (Pa s), r is the radius of the

spherical particle (m), and CW is the viscous drag coefficient due to the surface of the wall [25].
Electromagnetophoresis of microparticles
The electromagnetophoretic force exerted on small beads,
such as glass, copper and zinc particles, was experimentally
demonstrated by Kolin [24] and has been the subject of a
number of experimental studies using permanent magnets or
electromagnets. For example, Karni et al. [26] and Masujima et al. [27] reported the effect of magnetic field on electrophoresis of ions. Kolin proposed a method for the electrophoretic fractionation of particles suspended in a conductive liquid combined with electromagnetophoretic rotation
[28, 29]. Kolin et al. [30] and Nalbandian et al. [31] demonstrated electromagnetophoretic separation of biological
cells. Kolin [24] and Rice et al. [32] calculated the velocity
of microscopic particles. However, the velocities of microscopic particles were not observed, only predicted.
In our laboratory, the electromagnetophoreric migration velocity of microparticles was measured for polystyrene (PS) particles (3, 15, 20 m diameters) suspended
in a 1 M KCl aqueous solution, in a closed cell (250
250 m inner section) using rare-earth magnets (Nd-Fe-B,
0.1~0.2 T) [33]. The applied current ranged from 50 A
to 300 A. In all measurements, PS particles, which were
insulators, migrated in the direction of electromagnetic
buoyancy. The observed migration velocity of PS particles with diameter of 20 m was about 6 m s1 under the
conditions of 250 A and 0.105 T. This result indicates
that PS particles behave like dielectrics.
We investigated the electromagnetophoresis of micrometer-sized particles in a flow system under high magnetic field (up to 10 T) using a superconducting magnet
[34]. Figure 8 illustrates the apparatus for the observation

Fig. 8 Experimental setup for the observation of electromagnetophoresis: a superconducting magnet; b current source; c ammeter; d video recorder; e monitor; f syringe pump; g light source;
h CCD camera; i optical microscope; j capillary flow cell

of electromagnetophoresis in a capillary flow system under the high magnetic field. A square fused-silica capillary cell with 100 100 m inner section and 2 cm long
was set in the homogeneous magnetic field of the superconducting magnet. Both edges of the fused-silica capillary cell were equipped with Ag | AgCl electrodes and connected by PTFE tube. Using this apparatus, electromagnetophoretic velocities of microparticles, such as polystyrene
(PS), carbon, yeast cells and red blood cells (RBCs), were
measured. Figure 9 shows the electromagnetophoretic behavior of a PS particle with 20 m diameter in a flowing
aqueous solution. The direction of the flow in Fig. 9 was
from the right-hand side to the left-hand side. When a cur-

Fig. 9 Typical electromagnetophoretic behavior of a polystyrene

particle in the flow. The arrows show the direction of the applied current. The diameter of the polystyrene particle was 20 m. A 1.0 M KCl
solution was used as a conductive medium. The magnetic field was
10 T and the current was 10 A. The flow velocity was 5 L h-1.
This picture was reconstructed from the images captured at a rate of
5 frames/s

Fig. 10 Electromagnetophoretic velocity of microparticles at various applied currents. The magnitude of the homogeneous magnetic field was 10 T. Plots were: a PS particles with 10 m diameter; b red blood cells; c PS particles with 6 m diameter; d yeasts;
e carbon particles with 6 m diameter

1698
Table 1 Electrical conductives of particle p and fluid determined
f from the electromagnetophoresis
Particle Diameter
(m)

f/S cm1 p(intrinsic)/ p(apparent)/

(S cm1)
(S cm1a)

PS
Carbon
Yeast
RBC

0.112
0.110
0.112
0.019

3, 6, 10, 15, 20
6, 20
7.2
9.4

0
727
2103
6103

0.01910.0084
0.06680.0080
0.06740.0078
0.01070.0011

Table 2 The desorption current of particles iD and the adsorption

force between a particle and fused-silica surface FA calculated Eq.
(5) using iD, the adsorption force normalised divided by the radius
of particles
Particle

Diameter/
m

iD/A

FA/pN

FAr-l/
Nm-1

PS
PS
Carbon

15
20
20

549
311
309

684
997
426

93.7
99.7
38.7

Calculated from observed migration velocity using Eq. (6)

rent was applied, the particle moved diagonally to the direction of the flow due to the effect of electromagnetic
buoyancy. Figure 10 shows the migration velocities of
particles at a fixed homogeneous magnetic field of 10 T.
The polystyrene, yeasts, RBCs and carbon particles all
acted as insulators, although the carbon particle is intrinsically conductive. The apparent conductivities of the microparticles calculated by Eq. 6 using the observed migration velocities are given in Table 1. The calculated conductivities of the particles were clearly different from the
intrinsic values. This suggests that the apparent conductivity reflects a surface conductivity of the particles rather
than the intrinsic conductivity.

force, could be desorbed by a reversed electromagnetophoretic force by switching the direction of the current.
The current required to desorb the particles was measured,
and the adsorption force was calculated using Eq. 5. Table 2
gives the average values of the desorbed currents, and the
adsorption forces of PS particles with 15 m and 20 m
diameters and carbon particles with 20 m diameters. The
values of adsorption force divided by the radius for the PS
particles are almost constant regardless of the size, as shown
in Table 2. Because the electrostatic repulsion force is
very small in high ionic strength media, the adsorption
force of PS particles is mainly governed by van der Waals
force, which is represented by the equation [36]
)$ U = $  ' 

Measurement of adsorption force of a particle

by electromagnetophoretic force
Electromagnetophoretic force could be used for the measurement of the adsorption force between the particles and
the wall of a fused-silica capillary [35]. Figure 11 shows a
schematic drawing of the adsorption-desorption behavior
of a particle acting under electromagnetophoretic force.
A particle, once adsorbed by the electromagnetophoretic

(7)

where A is the Hamaker constant (J) and D is the distance

between the wall surface and the particle surface (m). The
present method is unique as a non-contacting adsorptionforce measurement method, with a high sensitivity of 1 pN.
This sensitivity is much higher than that of the ordinary
AFM method, which is typically 10 pN [37].
This desorption-adsorption control method using the
electromagnetophoretic force can now be applied to new
chromatographic separation techniques for microparticles
[38].

References

Fig. 11 Schematic illustration of the adsorption-desorption control of a particle by electromagnetophoretic force by switching the
direction of the applied current

1. Giddings JC, Brantley SL (1984) Separ Sci Technol 19:631

651
2. Vickrey TM, Garcia-Ramirez JA (1980) Separ Sci Technol 15:
12971304
3. Schunk TC, Gorse J, Burke MF (1984) Separ Sci Technol 19:
653666
4. Levin S, Giddings JC (1991) J Chem Technol Biot 50:4356
5. Fuh CB (2000) Anal Chem 72:266A-271A
6. Fuh CB, Chen SY (1998) J Chromatogr A 813:313324
7. Fuh CB, Chen SY (1999) J Chromatogr A 857:193204
8. Fuh CB, Lai JZ, Chang CM (2001) J Chromatogr A 923:263270
9. Chalmers JJ, Zborowski M, Sun L, Moore L (1998) Biotechnol
Progr 14:141148
10. Zborowski M, Sun L, Moore LR, Williams PS, Chalmers JJ
(1999) J Magn Magn Mater 194:224230
11. Williams PS, Zborowski M, Chalmers JJ (1999) Anal Chem
71:37993807
12. Moore LR, Rodriguez AR, Williams PS, McCloskey K, Bolwell BJ, Nakamura M, Chalmers JJ, Zborowski M (2001) J
Magn Magn Mater 225:227284
13. Watarai H, Namba M (2001) Anal Sci 17:i169-i171
14. Watarai H, Namba M (2002) J Chromatogr A 961:38

1699
15. Reddy S, Moore LR, Sun L, Zborowski M, Chalmers JJ (1996)
Chem Eng Sci 51:947956
16. Chalmers JJ, Zhao Y, Nakamura M, Melnik K, Lasky L, Moore
L, Zborowski M (1999) J Magn Magn Mater 194:231241
17. McCloskey KE, Comella K, Chalmers JJ, Margel S, Zborowski
M (2001) Biotechnol Bioeng 75:642655
18. McCloskey KE, Moore LR, Hoyos M, Rodriguez A, Chalmers
JJ, Zborowski M (2003) Biotechnol Progr 19:899907
19. Chosy EJ, Nakamura M, Melnik, Comella K, Lasky LC,
Zborowski M, Chalmers JJ (2003) Biotechnol Bioeng 82:
340351
20. Wilhelm C, Gazeau F, Bacri JC (2002) Eur Biophys J 31:118125
21. Suwa M, Watarai H (2001) Anal Chem 73:52145219
22. Suwa M, Watarai H (2002) Anal Chem 74:50275032
23. Suwa M, Watarai H (2003) J Chromatogr A (in press)
24. Kolin A (1953) Science 117:134137
25. Happel J, Benner H (1965) Low Reynolds number hydrodynamics. Prentice-Hall, Englewood Cliffs, NJ

26. Karni Z, Kopito LE (1975) Med Biol Eng 13:457462

27. Razee S, Tamura A, Masujima T (1994) Chem Pharm Bull 42:
23762378
28. Kolin A (1967) J Chromatogr 26:164179
29. Kolin A (1967) J Chromatogr 26:180193
30. Kolin A, Kado RT (1958) Nature 182:510512
31. Nalbandian RM, Michel RE, Mader I (1968) Experientia 24:
10061007
32. Rice CL, Whitehead R (1964) J Chem Phys 41:2202
33. Namba M, Watarai H (2000) Anal Sci 16:5-9
34. Iiguni Y, Watarai H (2003) J Chromatogr A (submitted)
35. Iiguni Y, Watarai H (2003) Anal Sci 19:3339
36. Israelachvili JN (1992) Intermolecular and surface forces, 2nd
edn. Academic, New York
37. Klein JD, Clapp AR, Dickinson RB (2003) J Colloid Interf Sci
261:379385
38. Iiguni Y, Watarai H (2003) Anal Chem (submitted)