Continuous
deviation
sorting
of magnetizable
particles
by means of specific
Roland Hartig and Michael Hausmann
Institute ofApplied Physics, University of Heidelberg, Albert-Ueberle-Strap
Germany
3-5, 69120 Heidelberg,
Georg Lijers
Institute of Anatomy and Cell Biology, University of Heidelberg, Im Neuenheimer Feld 307, 69120
Heidelberg
Michael Kraus
blstitute of Applied Physics, University of Heidelberg, Albert-Ueberle-Strap
Germany
3-5, 69120 Heidelberg,
Gerhard Weber
Dr. Meber GmbH, Am Englischen Garten 6, 85737 Ismaning, Germany
Christoph Cremer
Institute of Applied Physics, University of Heidelberg, Albert-Ueberle-Strafie
Germany
3-5, 69120 Heidelberg,
(Received 31 March 1994; accepted for publication 17 January 1995)
A device is described that has been developed for continuous magnetic sorting of particles which are
typically several micrometers in diameter. In a continuously streaming fluid, strongly magnetizable
particles are deviated in a different way from weakly magnetizable particles. The strongly
magnetizable particles are focused into a separate stream. The two types of particles can be collected
continuously into several vials. Sorting results and applications of this technique are
reported. 0 1995 American Institute of Physics.
1. INTRODUCTION
Magnetic separation is a widespread technique that has
been applied in commercial metal extraction and different
cases of material and life sciences. Different sorting devices
have been developed according to their respective requirements.‘” These techniques appear to be useful to handle
large quantities of particles. The differences in magnetic
properties, however, of biological particles usually are not
large enough to separate subpopulations of them. Thus magnetic solid supports with specific affinity couples have become a commonly used method for separating cells, cell organelles, and micro-organisms in biology and medicine.6
Magnetic microspheres can be labeled by antibodies. These
microspheres are polystyrene beads typically 1-5 ,%rn in diameter containing iron oxide.7>8The biological particles are
immunochemically fixed to these “magnetic beads” by a
highly specific antigen-antibody bridge.
Separation of the magnetic beads carrying biological
particles from nonmagnetic material can be performed by
appropriate devices; e.g., in some cases, a simple hand-held
permanent magnet may be used. This method appears to be
useful in cases of low density of magnetic beads and limited
requirements of sorting purity. Both components of the
sample, magnetically labeled and nonlabeled particles, are
homogeneously suspended in a separation vial. If the sample
is exposed to the magnet, the strongly magnetizable magnetic beads are forming chains, oriented into the direction of
the magnetic field. These chains are deviated into the direction of the magnetic force, i.e., in the direction of the magnetic field of the permanent magnet. Because nonlabeled particles are randomly dispersed in the separation volume, some
Rev. Sci. Instrum.
66 (5), May 1995
of these particles are in the region of the highest magnetic
forces at the beginning of the separation process. Into this
region the magnetically labeled particles are collected and
encIose those remaining nonmagnetic particles. This reduces
the sorting purity. Collective effects can induce agglomerations of particles’ that cannot be resuspended by washing
steps without impairing the conditions of preservation of the
biological particles.
Large numbers of particles can also be separated in magnetic filters.g’10 The filter material consists of magnetic stainless-steel wool in a magnetic field of 0.1-1.0 T. Nonlabeled
particles pass through the filter chamber, while magnetic microspheres and magnetically labeled particles are retained.
These can be flushed out when the magnetic field is switched
off. AIthough magnetic filtration has an indisputable success
in various separation technologies in biology and medicine,
there are still several basic aspects, especially in celI separation, which require further research. One disadvantage of
magnetic filtration is that the capacity of the filter chamber
must be coordinated with the number of particles retained in
the chamber. Too many particles may block the filter system
or lead to unwanted contamination in the separated fractions.
In the case of rare particles to be sorted out, there may be
significant deficit of particles. Another disadvantage is that
the retained particles can only be eluted by switching off the
magnetic field; continuous separation of large quantities of
particles cannot be performed. Additionally, both components of the sample are randomly dispersed in the whole
filter chamber. The disadvantage of such methods was discussed above. Retained cells rosetted by several magnetic
beads may be destroyed by forces of shear stress, induced by
0034-6740/95/66(5)/3289/7/$6.00
0 1995 American
Institute
of Physics
3289
Downloaded 28 Sep 2005 to 192.58.150.41. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
the strong magnetic field along the steel wool. These effects
can influence the vital sorting conditions of the retained particles.
Although there are different magnetic sorting devices
commercially available, several applications in medical research and routine application still require a high throughput
rate and a high purity combined with good vital conditions of
the biological particles that are sorted out. Most sorting devices do not fulfill these requirements simultaneously.
In this report, a device is described which allows continuous magnetic sorting by specific deviation of biological
particles labeled to magnetic beads that are some micrometers in diameter.
II. SEPARATION PROCESS
For sorting, the suspension containing magnetically labeled and nonlabeled particles is injected by a peristaltic
pump into a separation cell via two sample inlets. A separation medium inside is flowing in vertical direction upwards
as a thin film. The conditions of a strictly laminar flow have
to be preserved. The velocity of this flow is controlled by a
second peristaltic pump. The sample, being transported in
two narrow streams by the separation medium, passes an
inhomogeneous magnetic field that is perpendicularly oriented to the direction of the flow. According to the magnetic
moments, induced by the magnetic field, the strongly magnetizable particles, magnetic beads, and magnetically labeled
particles are deviated into the direction of the magnetic
forces, while the nonlabeled, weakly magnetizable particles
pass the magnetic field with negligible interaction. Depending on the shape of the poles of the electromagnet the deviated particles are focused into an area where the deviating
magnetic forces are equal to zero. Magnetic beads and the
magnetically labeled particles leave the area of the magnetic
field in a stream that is completely separated from the two
streams of the nonlabeled particles. After passing the interaction area of the magnetic field, the thin film of the separation medium is splitted into several fractions by a fraction
line, made of 60 Teflon tubes. These fractions are collected
into separate vials. The labeled and the nonlabeled particles
remain in suspension during the whole separation process.
The system has been developed for sorting applications in
biology and clinical medicine, for instance, vital sorting of
cells.ll
111.APPARATUS
The approach described above has been realized by an
apparatus consisting of two basic components, (A) the magnet configuration and (B) the separation chamber for the
streaming fluid transporting the sample through the magnetic
field.
A. Magnet configuration
An emerging technique to perform continuous magnetic
sorting is open-gradient separation. This technique deflects
’magnetizable particles into a spectrum based on the susceptibilities and volumes of the particles. One of the earliest
devices using open-gradient technology is the Frantz isody3290
Rev. Sci. Instrum.,
Vol. 66, No. 5, May 1995
X
*
FIG. 1. Field lines in the z-x plane and schematic diagram of the flowing
chamber between the two pole pieces of the magnet. The arrows are illustrating the directions of the deviating magnetic forces: 1: magnetic pole
pieces, 2: safety glass plate, 3: spacer gasket, 4: stream of the suspension of
the probe, and 5: stream of the focused particles (streams 4 and 5 are perpendicular to the drawing plane).
namic magnetic separator.’ This consists of two opposing
magnetic pole pieces producing a uniform separating force in
the air gap between the pole pieces. Varying the geometry of
the pole pieces, open-gradient technology can also be used to
perform continuous magnetic separation, by magnetic focusing of the deviated particles.
To realize an appropriate field configuration, a pair of
wedge-type magnetic pole pieces is used as shown in Fig. 1.
The configuration of the separation chamber and the magnetic-field lines between the pole pieces is also shown in this
figure. The field lines were calculated by a computer using a
program called Poisson.t’
The magnet configuration is schematically shown in Fig.
2. It consists of an electromagnet with a coil (160X 160X280
mm3) of 1000 turns of an isolated copper wire (2X3 mm2).
The magnetic induction can be varied by the current of the
coil. Up to a maximum current of 9 A, air cooling by means
of 10 mm thick air spaces between each of the seven turns of
the coil is sufficient.13 Inside the coil there is a magnetic core
(30X90 mm) of ST37 steel. The same material is used for
the two pole pieces (Fig. 2). These have a width of 40 mm (x
direction), a length of 300 mm (y direction) and are positioned with an air gap of 30 mm (z direction). The angles 0
of the pole tips are 120” as shown in Fig. 1.
The separation force on a single magnetizable particle in
a magnetic field with magnetic induction B=hH
is given
by r4
F= 112,uuxV,,V(H2),
Magnetizable
particles
Downloaded 28 Sep 2005 to 192.58.150.41. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
PIG. 2. Cross section of the magnet. The left figure demonstrates the z-s
cross section of the whole magnet. The right figure shows the z-y view of
the magnetic pole pieces to illustrate the form of the hyperbola. 1: coil, 2:
steel core.
where VP is the volume and x the effective susceptibility of
the particle. The direction of the magnetic force is determined by the direction of the gradient of the magnetic field.
If a particle is suspended in a fluid of susceptibility xf, the
force on the particle depends on the susceptibility difference
x of the particle x, and the surrounding fluid (x=xP--xf).
With respect to Eq. (l), the force on a magnetizable particle in the pole axis x between two wedge-type pole pieces
is given by14
F,-1
cot”$7
/Lo~xv~B2d2~
(2)
where d is the distance between the two poles pieces, B
the magnetic induction in the air gap, and 4~ an angle
[q~=0.5(180”-19); (&angle as shown in Fig. l)].
Particle displacement in the x direction is due only to the
magnetic forces and the viscous drag forces of the separation
medium in the chamber. Figure 3 shows the product of the
magnetic induction and its gradient (B dB/dx) as a function
of the coordinate x using an air gap of 30 mm and a magnetic
BdB/dx F%i-'l
_ _________________________
‘-l?t---
_________
r ______.____.
0
1
\
PIG. 4. Schema of the sorting chamber: The left figure presents the whole
chamber; the right figure shows the side view of the chamber and illustrates
the principle of the counter-flow technique. The arrows indicate the flow
directions. Additionally, iu the left figure the traces of the streams of the
different particles are shown. The streams of the weakly magnetizable nonlabeled particles are characterized by the two dotted lines. The stream of the
deviated particles is outlined by the dashed line. 1: front plate, 2: back plate,
3: spacer gasket, 4: sample inlet, 5: chamber buffer, 6: counter flow, 7:
fraction tube, and 8: fraction vial.
induction of 75 mT resulting from a current of 5 A in the coil
wires. The parameters of the magnetic field were measured
by a Hall element. In addition, the figure indicates that the
vectors of the deviating magnetic forces are directed antiparallel around the tip of the wedges of the two pole pieces. In
the region of the two extrema of the deviating forces, the
particle suspension passes this magnetic field. Strongly magnetizable particles are deviated into the direction of the gradient of the magnetic field. In the region of the tips of the
magnetic pole pieces, the deviating forces are equal to zero.
The strongly magnetizable particles are focused into one
stream that is located around the tips of the pole pieces of the
magnet. Weakly magnetizable particles pass the magnetic
field in two parallel streams without significant interaction.
Therefore the stream of the deviated strongly magnetizable
particles and the magnetically labeled biological particles is
completely separated from the streams of the nonlabeled particles (Fig. 4).
It is anticipated that if the magnetizable particles are
leaving the magnetic field, an additional gradient of the magnetic field that is directed antiparallel to the flow direction (y
direction) may stop the streaming particles. This would inhibit continuous separation. To overcome this disadvantage
this gradient must be as small as possible. A hyperbolic
shape of the pole pieces parallel to the direction of the flow
can make this problem negligible (Fig. 2).
With respect to Eq. (1) the force on a magnetizable particle in the symmetric axis of two pole pieces with such
hyperbolic shapes is given byl’
-oo5iL[mm,
30
-20
-10
0
10
20
30
PIG. 3. Plot of the product of the magnetic induction B and the gradient of
the induction dB/dx along the x axis. The current of the coil was fixed at 5.0
A during the measurement of the magnetic parameters.
Rev. Sci. Instrum.,
Vol. 66, No. 5, May 1995
F= - -$Vpc’kz(k-
l)r(2k-3),
Magnetizable
particles
3291
Downloaded 28 Sep 2005 to 192.58.150.41. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
where k is the numerical eccentricity of the hyperbola and c
a constant that depends on the magnetic induction B and the
distance of the pole pieces, i.e., the relevant dimension of the
air gap. r is a distance vector in the direction of the negative
y direction. For k = 1.5, a constant force along r can be obtained. In the case of the device described here, VB’ was
measured to be 0.02 T2 m-l. The resulting force is directed
antiparallel to the flow direction in the chamber and can be
compensated by the streaming forces of the fluid using an
appropriate buffer velocity.
number of particles per volume
800
400
0
16
B. Separation chamber
The chamber containing the flowing buffer medium
transporting the suspension of the sample through the magnetic field consists of two parallel safety glass plates 100
X640 mm2 (Fig. 4) spaced 0.5 mm apart. This distance is
given by a gasket, forming a leak-free seal. The separation
volume is 50 mm wide and 600 mm high. The buffer, injected by the buffer inlet (Fig. 4), passes continuously
through this chamber against the gravity force. 40 mm above
the buffer inlet, the sample inlets made of two Teflon tubes
are fixed into the front safety glass plate (Fig. 4). Convection
and diffusion are negligible under the experimental conditions chosen. The separated streams of particles are collected
in fractions by a counter-flow technique as shown in Fig. 4.
This means, that simultaneously to the chamber buffer, a
second buffer stream is injected by the counter-flow inlet.
This stream flows antiparallel to the chamber buffer with a
given velocity, which is controlled by a peristaltic pump. At
the array of the fraction tubes, the two buffer streams are
arranging in layers and the fluid is collected into 60 fractions.
An equal flow in all fraction tubes is necessary to fractionate
the different streams of particles with appropriate selectivity.
These flow conditions can be achieved by a flow velocity u
in a single tube of
AuR’
where R is the radius of the tube (0.15 mm), I the length of
the tube (500 mm), 17the dynamic viscosity of the fluid, and
Ap the difference of pressure that is given by the altitude of
the fraction tubes as shown in Fig. 4. For example, using
water to be the buffer medium and an altitude of 300 mm
between the fraction line and fraction tube, these parameters
result in v 31.7X lo-” m/s. This value defines the ratio of the
flow velocity between the chamber buffer and the buffer of
the counter flow. A very small velocity of flow of the chamber buffer must be compensated by a very high velocity of
the counter tlow. The z profile of the streaming fluid is defined by the velocity of the fluid and the z width of the
separation volume. This z profile determines the greatest
length of the chains of magnetizable particles which are induced by the magnetic field. As shown in Fig. 1, this is also
oriented in z direction. Therefore the z width of the separation volume was limited to 0.5 mm, defined by the gasket.
3292
Rev. Sci. Instrum.,
Vol. 66, No. 5, May 1995
IO
1-T
14
18
22
26
30
action
number
FIG. 5. Sorting result after magnetic focusing of test particles. The deviated
stream of the strongly magnetizable particles 5 Cy=O.13) is completely
separated from the nondeviated streams of the weakly magnetizable particles Cl (x=1.2X lo+).
IV. EXPERIMENTAL RESULTS AND DISCUSSION
To test the sorting device, weakly magnetizable particles,
‘<latex beads” (,Y,= 1.2X 10e6, diameter=4.3 ,um), and
strongly magnetizable particles, magnetic beads &=0.13,
diameter=4.5 pm), were mixed in different concentrations
and sorted: the particles were injected into the sorting chamber by the two sample inlets at a rate of 5.0 ml per hour each,
and passed the magnetic field in two parallel streams. The
flow rate of the separation medium (0.15 M NaCl, 0.05%
serum albumin bovine, pH 7.4) was fixed to 2.1 X 1O-3 m/s.
The current of the electromagnet was adjusted to 5.0 A, inducing a magnetic field of 75 mT in the air gap between the
two wedge-type pole pieces. Magnetic beads were deviated
and focused into a stream localized between the two parallel
particle traces of the weakly magnetizable latex beads. In
Fig. 5 a typical sorting result is shown. In this experiment,
the suspension of test particles was injected at a density of
3X lo7 particles per milliliter under the given experimental
conditions. The ratio between the latex beads and the magnetic beads was 2~1. The figure shows the counted numbers
of particles per given volume of the fractions as determined
by microscopic evaluation. In the nondeviated fractions
(fraction numbers 7-9 and 28-30) more than 99.5% of the
beads were found to be latex beads and in the deviated fractions (fraction numbers 15-22) more than 99.5% of the
beads were found to be magnetic beads. For each fraction
(shown in Fig. 5) at least 1000 particles were counted. Therefore from its physical properties, specific deviation of magnetizable particles offers sorting purities of more than 99%.
To test the capabilities of the sorting system, suspensions of
latex and magnetic beads in different concentrations and ratios were used. The ratio between the number of weakly
magnetizable latex beads and the strongly magnetizable
magnetic beads was varied from 1OOO:l up to 1:400. The
concentration of the suspension differed between 1 X lo6 and
1 X lo8 particles per milliliter. In all experiments sorting purities were obtained in order of 99% under the conditions of
a high throughput rate. Within 5 min up to 8X107 particles
were sorted out corresponding to a throughput rate of about
IX lo9 particles per hour. Counting the number of the parMagnetizable
particles
Downloaded 28 Sep 2005 to 192.58.150.41. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
(a)
FLl-tl\FLi-Height
--->
(c)
FLi-H\FLl-Height
--->
FIG. 6. Flow cytometry plot: PEfluorescence (FL2) vs FITC-fluorescence
(FLl) for B- and T-cell detection. Each dot represents a cell with a given
PE/FITC fluorescence. The B-cells were stained by phycoerythrin (PE)labeled anti-CD19 antibodies and the T-cells were stained by fluorescein
isothiocyanate @TO-labeled anti-CD2 antibodies (CDZFITC). (a) Before
magnetic sorting: most cells were T-cells; (b) magnetically deviated fractions after magnetic sorting: the B-cells were highly enriched; (c) nondeviated fractions: nearly all B-cells were depleted.
(b)
FLl-H\FLl-Height
--->
titles in aliquots of the nondeviated as well as of the deviated
fractions and comparing this number with the number of
particles in the suspension before magnetic sorting, a recovery of more than 98.5% was detected.
However, in contrast to the results shown here it. has to
be reminded that for applications in biology, for instance,
vital cell sorting or chromosome sorting, the inherent magnetic properties of the biological material usually cannot be
used. They have to be specifically attached to magnetic beads
as described above.6
Magnetic labeling of biological particles can result in
three different experimental situations: (a) several magnetic
beads are bound to one biologicaLparticle, (b) one magnetic
bead is bound to one biological pazicle, (c) several biological particles are immunochemically fixed to one magnetic
bead. The ratio between the diameter of the magnetic beads
and the diameter of the biological particles defines the relative amount of these particles.
Application (a) was performed by sorting out a subpopulation (B-cells) of white blood cells (WBC), a standard problem of immunology. The WBC were enriched from human
peripheral blood by density centrifugation (Ficoll method). A
Rev. Sci. Instrum.,
Vol. 66, No. 5, May 1995
subpopulation was labeled by magnetic beads (Dynabeads
M-450, Dynal, Norway) which were 4.5 pm in diameter and
which were precoated with CD 19 antibodies directed
against B-lymphocytes. The sample was injected to the sorting device at a rate of 4X108 particles per hour. The sorting
conditions were the same as described above, performing
separation experiments of beads only. Aliquots of the sample
before sorting and the sorting fractions were analyzed by
flow cytometry.16 Before magnetic sorting the sample contained 9.4% B-lymphocytes [Fig. 6(a)]. After magnetic sorting, 85% of the cells in the deviated fractions were identified
as B-cells [Fig. 6(b)]. The nondeviated fractions contained
less than 0.1% B-lymphocytes [Fig. 6(c)]. For each flow cytometry measurement, 5000 particles were acquired. The viability of the cells of the sorting fractions was tested by flow
cytometry following propidium iodide staining of’the cells.17
It resulted in a rate of more than 99% viable cells. This
shows that the sorting device offers the possibility ,of viable
celI sorting. Comparing these results of magnetic sorting to
the experiments with isolated beads, the respective purity of
the deviated fractions considerably differed from the technical optimum. This may be due to the nature of such cells,
Magnetizable
particles
3293
Downloaded 28 Sep 2005 to 192.58.150.41. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
which are likely to agglomerate. The agglomerations also
containing unspecific material were magnetically deviated if
they included cells labeled with antibodies and hence with
magnetic beads. This disadvantage limited the purity of the
deviated fractions but may be overcome by improved preparation techniques.
In experiments with isolated chromosomes in suspension, application (.b) was performed. Metaphase chromosomes were isolated from cultured cells of the Chinese hamster. One chromosome (length GlO pm) was bound to one
magnetic bead, which was 4.5 ,um in diameter (Dynabeads
M-450, Dynal, Norway). For labeling, DNA specific antibodies were used (Antibody AC30-10, Progen Heidelberg, Germany) directed against the sugar-phosphate backbone of
single and double stranded DNA of the chromosomes. This
procedure labels all chromosomes and may be a helpful tool
to purify chromosome suspensions from cell debris. Furthermore, preliminary experiments using in situ hybridization
techniques’s were performed with chromosomes in suspension. The specificity that can be obtained by such techniques
is known from fluorescence labeling of chromosomes fixed
on slides for microscopic analysis.‘9-2* Thus the magnetic
separation method described here also offers new possibilities for the bulk sorting of chromosomes in suspension. Here
we report some basic experiments to show the technical capacities of the sorting device.
isolated chromosomes in suspension were magnetically
labeled and subjected to the sorting device and separated
according to application schema (b) described above. After
magnetic sorting and fluorescence staining, the chromosomes
still showed a well-preserved morphology as detected by
light microscopy. In the deviated fractions, 99% of the particles were found to be magnetically labeled chromosomes or
magnetic beads. In the nondeviated fractions less than 0.1%
strongly magnetizable particles (magnetic beads and labeled
chromosomes) were detected. For each measurement 1000
particles were inspected microscopically.
Since the separation technique presented here uses welldefined deviation forces, the sorting process also allows studies of the mechanical stability of the binding bridge between
the magnetic beads and the labeled biological particles. Preliminary results of such studies, performing different labeling
techniques of chromosomes (antibody binding and in situ
hybridization), revealed a difference between the strength of
the binding bridge of specifically labeled chromosomes and
of unspecific attachment. To perform these experiments, the
sample was injected at a rate of 4.5 ml/h into the separation
chamber by one sample inlet only. Additionally, the deviation
distance was fixed to 15 mm by displacing the tips of the two
pole pieces relative to the sample stream (Fig. lj. The velocity of the separation medium was fixed to 2.1 X10-s m/s. All
chromosomes unspecifically attached to magnetic beads
were detached during the sorting process, using deviation
forces of about lo-r2 N: within the separation distance the
attached chromosomes were carried off the magnetic beads
by the forces of shear stress that were induced by the hydrodynamic forces around the deviated particles. Only magnetic
beads without any chromosomes were found in the deviated
fractions. The binding bridge of the specifically labeled chro3294
Rev. Sci. Instrum.,
Vol. 66, No. 5, May 1995
mosomes, however, was maintained as detected by light microscopy. Thus magnetic sorting by specific deviation allowed us to discriminate between unspecsc attachment and
specific labeling.
For application (c), experiments on the immunomagnetic
isoIation of cell organelles (peroxisomes of rat liver, 150500 nm diameter) were performed. After coating magnetic
beads that were 2.8 and 4.5 pm in diameter (Dynabeads
M-280 and M-450, Dynal, Norway) with suitable antibodies,
up to 20 organelles were bound to one magnetic bead. Scanning electron microscopy revealed the decoration of the
beads. After separation, the organelles still showed a wellpreserved morphology as detected by transmission electron
microscopy. The purities of the deviated fractions were only
limited by the specificity of the antibodies used in these experiments. For biochemical analysis, the separated organelles
were detached from the magnetic beads by adding a glycin
buffer (0.2 M, pH 2.8) to the pelleted particles of the deviated fractions. The results of these experiments indicated that
magnetic separation by specific deviation is an alternative
approach for a reliable and specific isolation of highly purified small cell organelles.”
The results of the basic experiments described above
showed that magnetic separation by means of specific deviation allows the separation of biological particles that were
from several hundreds .of nanometers up to several tens of
micrometers in diameter. These separations were performed
at a high throughput rate. Simultaneously, reasonable high
sorting purities were achieved. In principle the sorting result
(purity, throughput rate) is limited not by the physical procedure itself but by biological and biochemical parameters
such as the specificity of the labeling technique as well as the
preparation conditions of the biological objects.23 Moreover,
it is of great advantage that the biochemical composition of
the chamber buffer is independent from the sorting process.
This allows optimal conditions to preserve the viability of
the biological particles, e.g., living cells. This may open new
aspects for many sorting problems in cell biology and medicine. Besides these applications, sorting of technically
relevant “nonbiological” particles can also be performed.
Changing the separation volume and the size of the magnetic
pole pieces and magnetic coil allows the separation of larger
particles by means of continuous magnetic focusing. Such
applications may be feasible in recycling technology.
Furthermore, particles with different susceptibilities and volumes [Eq. (l)] can be separated continuously in “free-fall”
arrangements, so that in certain cases no chamber buffer at
all will be necessary (Ref. 24 and our own unpublished results).
ACKNOWLEDGMENTS
The authors acknowledge the computer support of the
German Cancer Research Center (DKFZ), Heidelberg, and
of the Max-Planck Institute for Nuclear Physics, Heidelberg.
The authors also wish to thank Dr. J. Schmitt, Immunopharmacology Department of Boehringer Mannheim GmbH,
Mamrheim, for performing flow cytometry measurements.
They also wish to acknowledge the support of the Deutsche
Forschungsgemeinschaft (DFG).
Magnetizable
particles
Downloaded 28 Sep 2005 to 192.58.150.41. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
‘J. Oberteuffer, IEEE Trans. Magn. 10, 223 (1974).
‘H. Kolm, J. Oberteuffer, and D. Kelland, Sci. Am. 11, 46 (1975).
3L. Petrakis, P. F. Abner, and F. B. Kiviat, Sep. Sci. Tech. 16, 745 (1981).
4B. B. Emroy, IEEE Trans. Magn. 18, 1686 (1982).
‘K. M. Jeong, L. Petrakis, M. Takayasu, and F. J. Friedlaender, IEEE Trans.
Magn. 18, 1692 11982).
“J. T. Kemshead and J. Ugelstad, Mol. Cell. Biochem. 67, 11 (1985).
7T. Lea, F. Vartdal, C Davies, and J. Ugelstad, &and. J. Immunol. 22, 207
(1985).
sJ. Ugelstad, L. Soderberg, A. Berge, and J. Bergstrom, Nature 303, 9.5
( 1983).
‘S. Miltenyi, W. Miiller, W. Weichel, and A. Radbruch, Cytometry 11, 231
(1990).
l”J. Plavins and M. Lauva, J. Magn. Magn. Mater. 122, 349 (1993).
“R Hartig, M Hausmann, J. Schmitt, D. B. J. Herrmann, M. Riedmiller,
and C. Cremer, Electrophoresis 13, 674 (1992).
*aA. M. Winslow, J. Comput. Phys. 2, 149 (1967).
Rev. Sci. Instrum.,
Vol. 66, No. 5, May 1995
r3G. Schnell, Magn&(Thiemig,
Miinchen, 1973), p. 101.
i4J. D. Jackson, Classical Electrodynamics (Wiley, New York, 1975).
“1. Yaniv, Y. Zimmels, and I. I. Lin, Sep. Sci. Tech. 14, 261 (1979).
I6 M. R. Melamed, T. Lindmo, and M. L Mendelsohn, Flow Cytometry and
Sorting (Wiley, New York, 1990).
“A. Krishan, J. Cell Biol. 66, 188 (1975).
“G. Dudin, E. W. Steegmayer, P. Vogt, H Schnitzer, E. Diaz, K. E. Howell,
T. Cremer, and C. Cremer, Hum. Genet. 80, 111 (1988).
r9P. Lichter, A. L. Boyle, T. Cremer, and D. C. Ward, Genet. Anal. Technol.
Appl. 8, 24 (1991).
*‘C. Cremer and T. Cremer, Eur. J. Histochem. 36, 15 (1992).
21D C Tkachuk, D. Pinkel, W.-L. Kuo, H.-U. Weier, and J. W. Gray, Genet.
AnaI. Technol. Appl. 8, 67 (1991).
“G. Liiers, R. Hartig, M. Hausmann, H. D. Fahimi, C. Cremer, and A.
VOlkl, Mol. Biol. Cell Suppl. 5, 476a (1994).
m R. Hartig, Thesis, University of Heidelberg, 1993.
“E. C. Hise, IEEE Trans. Magn. 18, 847 (1982).
Magnetizable
particles
3295
Downloaded 28 Sep 2005 to 192.58.150.41. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp